ngskcloud
/
gcgj-dify-1.7.0
17
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Merge branch 'langgenius:main' into main

Browse Source
pull/19119/head
sondin 1 year ago committed by GitHub
parent 0ee504ec64 8bf3f5ea78
commit e9599876a9
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
352 changed files with 2340 additions and 1811 deletions
Show Stats Download Patch File Download Diff File

2
.devcontainer/README.md
Unescape Escape View File

@ -34,4 +34,4 @@ if you see such error message when you open this project in codespaces:
![Alt text](troubleshooting.png)
a simple workaround is change `/signin` endpoint into another one, then login with GitHub account and close the tab, then change it back to `/signin` endpoint. Then all things will be fine.
The reason is `signin` endpoint is not allowed in codespaces, details can be found [here](https://github.com/orgs/community/discussions/5204)
The reason is `signin` endpoint is not allowed in codespaces, details can be found [here](https://github.com/orgs/community/discussions/5204)

2
.devcontainer/devcontainer.json
Unescape Escape View File

@ -2,7 +2,7 @@
// README at: https://github.com/devcontainers/templates/tree/main/src/anaconda
{
"name": "Python 3.12",
"build": {
"build": {
"context": "..",
"dockerfile": "Dockerfile"
},

4
.devcontainer/noop.txt
Unescape Escape View File

@ -1,3 +1,3 @@
This file copied into the container along with environment.yml* from the parent
folder. This file is included to prevents the Dockerfile COPY instruction from
failing if no environment.yml is found.
folder. This file is included to prevents the Dockerfile COPY instruction from
failing if no environment.yml is found.

25
web/.editorconfig → .editorconfig
Unescape Escape View File

@ -5,18 +5,35 @@ root = true
# Unix-style newlines with a newline ending every file
[*]
charset = utf-8
end_of_line = lf
insert_final_newline = true
trim_trailing_whitespace = true
[*.py]
indent_size = 4
indent_style = space
[*.{yml,yaml}]
indent_style = space
indent_size = 2
[*.toml]
indent_size = 4
indent_style = space
# Markdown and MDX are whitespace sensitive languages.
# Do not remove trailing spaces.
[*.{md,mdx}]
trim_trailing_whitespace = false
# Matches multiple files with brace expansion notation
# Set default charset
[*.{js,tsx}]
charset = utf-8
indent_style = space
indent_size = 2
# Matches the exact files either package.json or .travis.yml
[{package.json,.travis.yml}]
# Matches the exact files package.json
[package.json]
indent_style = space
indent_size = 2

2
.gitattributes vendored
Unescape Escape View File

@ -1,5 +1,5 @@
# Ensure that .sh scripts use LF as line separator, even if they are checked out
# to Windows(NTFS) file-system, by a user of Docker for Windows.
# to Windows(NTFS) file-system, by a user of Docker for Windows.
# These .sh scripts will be run from the Container after `docker compose up -d`.
# If they appear to be CRLF style, Dash from the Container will fail to execute
# them.

22
.github/linters/editorconfig-checker.json vendored
Unescape Escape View File

@ -0,0 +1,22 @@
{
"Verbose": false,
"Debug": false,
"IgnoreDefaults": false,
"SpacesAfterTabs": false,
"NoColor": false,
"Exclude": [
"^web/public/vs/",
"^web/public/pdf.worker.min.mjs$",
"web/app/components/base/icons/src/vender/"
],
"AllowedContentTypes": [],
"PassedFiles": [],
"Disable": {
"EndOfLine": false,
"Indentation": false,
"IndentSize": true,
"InsertFinalNewline": false,
"TrimTrailingWhitespace": false,
"MaxLineLength": false
}
}

17
.github/workflows/style.yml vendored
Unescape Escape View File

@ -9,6 +9,12 @@ concurrency:
group: style-${{ github.head_ref || github.run_id }}
cancel-in-progress: true
permissions:
checks: write
statuses: write
contents: read
jobs:
python-style:
name: Python Style
@ -163,3 +169,14 @@ jobs:
VALIDATE_DOCKERFILE_HADOLINT: true
VALIDATE_XML: true
VALIDATE_YAML: true
- name: EditorConfig checks
uses: super-linter/super-linter/slim@v7
env:
DEFAULT_BRANCH: main
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
IGNORE_GENERATED_FILES: true
IGNORE_GITIGNORED_FILES: true
# EditorConfig validation
VALIDATE_EDITORCONFIG: true
EDITORCONFIG_FILE_NAME: editorconfig-checker.json

2
CONTRIBUTING_ES.md
Unescape Escape View File

@ -90,4 +90,4 @@ Recomendamos revisar este documento cuidadosamente antes de proceder con la conf
No dudes en contactarnos si encuentras algún problema durante el proceso de configuración.
## Obteniendo Ayuda
Si alguna vez te quedas atascado o tienes una pregunta urgente mientras contribuyes, simplemente envíanos tus consultas a través del issue relacionado de GitHub, o únete a nuestro [Discord](https://discord.gg/8Tpq4AcN9c) para una charla rápida.
Si alguna vez te quedas atascado o tienes una pregunta urgente mientras contribuyes, simplemente envíanos tus consultas a través del issue relacionado de GitHub, o únete a nuestro [Discord](https://discord.gg/8Tpq4AcN9c) para una charla rápida.

2
CONTRIBUTING_FR.md
Unescape Escape View File

@ -90,4 +90,4 @@ Nous recommandons de revoir attentivement ce document avant de procéder à la c
N'hésitez pas à nous contacter si vous rencontrez des problèmes pendant le processus de configuration.
## Obtenir de l'aide
Si jamais vous êtes bloqué ou avez une question urgente en contribuant, envoyez-nous simplement vos questions via le problème GitHub concerné, ou rejoignez notre [Discord](https://discord.gg/8Tpq4AcN9c) pour une discussion rapide.
Si jamais vous êtes bloqué ou avez une question urgente en contribuant, envoyez-nous simplement vos questions via le problème GitHub concerné, ou rejoignez notre [Discord](https://discord.gg/8Tpq4AcN9c) pour une discussion rapide.

2
CONTRIBUTING_KR.md
Unescape Escape View File

@ -90,4 +90,4 @@ PR 설명에 기존 이슈를 연결하거나 새 이슈를 여는 것을 잊지
설정 과정에서 문제가 발생하면 언제든지 연락해 주세요.
## 도움 받기
기여하는 동안 막히거나 긴급한 질문이 있으면, 관련 GitHub 이슈를 통해 질문을 보내거나, 빠른 대화를 위해 우리의 [Discord](https://discord.gg/8Tpq4AcN9c)에 참여하세요.
기여하는 동안 막히거나 긴급한 질문이 있으면, 관련 GitHub 이슈를 통해 질문을 보내거나, 빠른 대화를 위해 우리의 [Discord](https://discord.gg/8Tpq4AcN9c)에 참여하세요.

2
CONTRIBUTING_PT.md
Unescape Escape View File

@ -90,4 +90,4 @@ Recomendamos revisar este documento cuidadosamente antes de prosseguir com a con
Sinta-se à vontade para entrar em contato se encontrar quaisquer problemas durante o processo de configuração.
## Obtendo Ajuda
Se você ficar preso ou tiver uma dúvida urgente enquanto contribui, simplesmente envie suas perguntas através do problema relacionado no GitHub, ou entre no nosso [Discord](https://discord.gg/8Tpq4AcN9c) para uma conversa rápida.
Se você ficar preso ou tiver uma dúvida urgente enquanto contribui, simplesmente envie suas perguntas através do problema relacionado no GitHub, ou entre no nosso [Discord](https://discord.gg/8Tpq4AcN9c) para uma conversa rápida.

2
CONTRIBUTING_TR.md
Unescape Escape View File

@ -90,4 +90,4 @@ Kuruluma geçmeden önce bu belgeyi dikkatlice incelemenizi öneririz, çünkü
Kurulum süreci sırasında herhangi bir sorunla karşılaşırsanız bizimle iletişime geçmekten çekinmeyin.
## Yardım Almak
Katkıda bulunurken takılırsanız veya yanıcı bir sorunuz olursa, sorularınızı ilgili GitHub sorunu aracılığıyla bize gönderin veya hızlı bir sohbet için [Discord'umuza](https://discord.gg/8Tpq4AcN9c) katılın.
Katkıda bulunurken takılırsanız veya yanıcı bir sorunuz olursa, sorularınızı ilgili GitHub sorunu aracılığıyla bize gönderin veya hızlı bir sohbet için [Discord'umuza](https://discord.gg/8Tpq4AcN9c) katılın.

518
README_SI.md
Unescape Escape View File

@ -1,259 +1,259 @@
![cover-v5-optimized](https://github.com/langgenius/dify/assets/13230914/f9e19af5-61ba-4119-b926-d10c4c06ebab)
<p align="center">
📌 <a href="https://dify.ai/blog/introducing-dify-workflow-file-upload-a-demo-on-ai-podcast">Predstavljamo nalaganje datotek Dify Workflow: znova ustvarite Google NotebookLM Podcast</a>
</p>
<p align="center">
<a href="https://cloud.dify.ai">Dify Cloud</a> ·
<a href="https://docs.dify.ai/getting-started/install-self-hosted">Samostojno gostovanje</a> ·
<a href="https://docs.dify.ai">Dokumentacija</a> ·
<a href="https://dify.ai/pricing">Pregled ponudb izdelkov Dify</a>
</p>
<p align="center">
<a href="https://dify.ai" target="_blank">
<img alt="Static Badge" src="https://img.shields.io/badge/Product-F04438"></a>
<a href="https://dify.ai/pricing" target="_blank">
<img alt="Static Badge" src="https://img.shields.io/badge/free-pricing?logo=free&color=%20%23155EEF&label=pricing&labelColor=%20%23528bff"></a>
<a href="https://discord.gg/FngNHpbcY7" target="_blank">
<img src="https://img.shields.io/discord/1082486657678311454?logo=discord&labelColor=%20%235462eb&logoColor=%20%23f5f5f5&color=%20%235462eb"
alt="chat on Discord"></a>
<a href="https://twitter.com/intent/follow?screen_name=dify_ai" target="_blank">
<img src="https://img.shields.io/twitter/follow/dify_ai?logo=X&color=%20%23f5f5f5"
alt="follow on X(Twitter)"></a>
<a href="https://www.linkedin.com/company/langgenius/" target="_blank">
<img src="https://custom-icon-badges.demolab.com/badge/LinkedIn-0A66C2?logo=linkedin-white&logoColor=fff"
alt="follow on LinkedIn"></a>
<a href="https://hub.docker.com/u/langgenius" target="_blank">
<img alt="Docker Pulls" src="https://img.shields.io/docker/pulls/langgenius/dify-web?labelColor=%20%23FDB062&color=%20%23f79009"></a>
<a href="https://github.com/langgenius/dify/graphs/commit-activity" target="_blank">
<img alt="Commits last month" src="https://img.shields.io/github/commit-activity/m/langgenius/dify?labelColor=%20%2332b583&color=%20%2312b76a"></a>
<a href="https://github.com/langgenius/dify/" target="_blank">
<img alt="Issues closed" src="https://img.shields.io/github/issues-search?query=repo%3Alanggenius%2Fdify%20is%3Aclosed&label=issues%20closed&labelColor=%20%237d89b0&color=%20%235d6b98"></a>
<a href="https://github.com/langgenius/dify/discussions/" target="_blank">
<img alt="Discussion posts" src="https://img.shields.io/github/discussions/langgenius/dify?labelColor=%20%239b8afb&color=%20%237a5af8"></a>
</p>
<p align="center">
<a href="./README.md"><img alt="README in English" src="https://img.shields.io/badge/English-d9d9d9"></a>
<a href="./README_CN.md"><img alt="简体中文版自述文件" src="https://img.shields.io/badge/简体中文-d9d9d9"></a>
<a href="./README_JA.md"><img alt="日本語のREADME" src="https://img.shields.io/badge/日本語-d9d9d9"></a>
<a href="./README_ES.md"><img alt="README en Español" src="https://img.shields.io/badge/Español-d9d9d9"></a>
<a href="./README_FR.md"><img alt="README en Français" src="https://img.shields.io/badge/Français-d9d9d9"></a>
<a href="./README_KL.md"><img alt="README tlhIngan Hol" src="https://img.shields.io/badge/Klingon-d9d9d9"></a>
<a href="./README_KR.md"><img alt="README in Korean" src="https://img.shields.io/badge/한국어-d9d9d9"></a>
<a href="./README_AR.md"><img alt="README بالعربية" src="https://img.shields.io/badge/العربية-d9d9d9"></a>
<a href="./README_TR.md"><img alt="Türkçe README" src="https://img.shields.io/badge/Türkçe-d9d9d9"></a>
<a href="./README_VI.md"><img alt="README Tiếng Việt" src="https://img.shields.io/badge/Ti%E1%BA%BFng%20Vi%E1%BB%87t-d9d9d9"></a>
<a href="./README_SI.md"><img alt="README Slovenščina" src="https://img.shields.io/badge/Sloven%C5%A1%C4%8Dina-d9d9d9"></a>
<a href="./README_BN.md"><img alt="README in বাংলা" src="https://img.shields.io/badge/বাংলা-d9d9d9"></a>
</p>
Dify je odprtokodna platforma za razvoj aplikacij LLM. Njegov intuitivni vmesnik združuje agentski potek dela z umetno inteligenco, cevovod RAG, zmogljivosti agentov, upravljanje modelov, funkcije opazovanja in več, kar vam omogoča hiter prehod od prototipa do proizvodnje.
## Hitri začetek
> Preden namestite Dify, se prepričajte, da vaša naprava izpolnjuje naslednje minimalne sistemske zahteve:
>
>- CPU >= 2 Core
>- RAM >= 4 GiB
</br>
Najlažji način za zagon strežnika Dify je prek docker compose . Preden zaženete Dify z naslednjimi ukazi, se prepričajte, da sta Docker in Docker Compose nameščena na vašem računalniku:
```bash
cd dify
cd docker
cp .env.example .env
docker compose up -d
```
Po zagonu lahko dostopate do nadzorne plošče Dify v brskalniku na [http://localhost/install](http://localhost/install) in začnete postopek inicializacije.
#### Iskanje pomoči
Prosimo, glejte naša pogosta vprašanja [FAQ](https://docs.dify.ai/getting-started/install-self-hosted/faqs) če naletite na težave pri nastavitvi Dify. Če imate še vedno težave, se obrnite na [skupnost ali nas](#community--contact).
> Če želite prispevati k Difyju ali narediti dodaten razvoj, glejte naš vodnik za [uvajanje iz izvorne kode](https://docs.dify.ai/getting-started/install-self-hosted/local-source-code)
## Ključne značilnosti
**1. Potek dela**:
Zgradite in preizkusite zmogljive poteke dela AI na vizualnem platnu, pri čemer izkoristite vse naslednje funkcije in več.
https://github.com/langgenius/dify/assets/13230914/356df23e-1604-483d-80a6-9517ece318aa
**2. Celovita podpora za modele**:
Brezhibna integracija s stotinami lastniških/odprtokodnih LLM-jev ducatov ponudnikov sklepanja in samostojnih rešitev, ki pokrivajo GPT, Mistral, Llama3 in vse modele, združljive z API-jem OpenAI. Celoten seznam podprtih ponudnikov modelov najdete [tukaj](https://docs.dify.ai/getting-started/readme/model-providers).
![providers-v5](https://github.com/langgenius/dify/assets/13230914/5a17bdbe-097a-4100-8363-40255b70f6e3)
**3. Prompt IDE**:
intuitivni vmesnik za ustvarjanje pozivov, primerjavo zmogljivosti modela in dodajanje dodatnih funkcij, kot je pretvorba besedila v govor, aplikaciji, ki temelji na klepetu.
**4. RAG Pipeline**:
E Obsežne zmogljivosti RAG, ki pokrivajo vse od vnosa dokumenta do priklica, s podporo za ekstrakcijo besedila iz datotek PDF, PPT in drugih običajnih formatov dokumentov.
**5. Agent capabilities**:
definirate lahko agente, ki temeljijo na klicanju funkcij LLM ali ReAct, in dodate vnaprej izdelana orodja ali orodja po meri za agenta. Dify ponuja več kot 50 vgrajenih orodij za agente AI, kot so Google Search, DALL·E, Stable Diffusion in WolframAlpha.
**6. LLMOps**:
Spremljajte in analizirajte dnevnike aplikacij in učinkovitost skozi čas. Pozive, nabore podatkov in modele lahko nenehno izboljšujete na podlagi proizvodnih podatkov in opomb.
**7. Backend-as-a-Service**:
AVse ponudbe Difyja so opremljene z ustreznimi API-ji, tako da lahko Dify brez težav integrirate v svojo poslovno logiko.
## Primerjava Funkcij
<table style="width: 100%;">
<tr>
<th align="center">Funkcija</th>
<th align="center">Dify.AI</th>
<th align="center">LangChain</th>
<th align="center">Flowise</th>
<th align="center">OpenAI Assistants API</th>
</tr>
<tr>
<td align="center">Programski pristop</td>
<td align="center">API + usmerjeno v aplikacije</td>
<td align="center">Python koda</td>
<td align="center">Usmerjeno v aplikacije</td>
<td align="center">Usmerjeno v API</td>
</tr>
<tr>
<td align="center">Podprti LLM-ji</td>
<td align="center">Bogata izbira</td>
<td align="center">Bogata izbira</td>
<td align="center">Bogata izbira</td>
<td align="center">Samo OpenAI</td>
</tr>
<tr>
<td align="center">RAG pogon</td>
<td align="center">✅</td>
<td align="center">✅</td>
<td align="center">✅</td>
<td align="center">✅</td>
</tr>
<tr>
<td align="center">Agent</td>
<td align="center">✅</td>
<td align="center">✅</td>
<td align="center">❌</td>
<td align="center">✅</td>
</tr>
<tr>
<td align="center">Potek dela</td>
<td align="center">✅</td>
<td align="center">❌</td>
<td align="center">✅</td>
<td align="center">❌</td>
</tr>
<tr>
<td align="center">Spremljanje</td>
<td align="center">✅</td>
<td align="center">✅</td>
<td align="center">❌</td>
<td align="center">❌</td>
</tr>
<tr>
<td align="center">Funkcija za podjetja (SSO/nadzor dostopa)</td>
<td align="center">✅</td>
<td align="center">❌</td>
<td align="center">❌</td>
<td align="center">❌</td>
</tr>
<tr>
<td align="center">Lokalna namestitev</td>
<td align="center">✅</td>
<td align="center">✅</td>
<td align="center">✅</td>
<td align="center">❌</td>
</tr>
</table>
## Uporaba Dify
- **Cloud </br>**
Gostimo storitev Dify Cloud za vsakogar, ki jo lahko preizkusite brez nastavitev. Zagotavlja vse zmožnosti različice za samostojno namestitev in vključuje 200 brezplačnih klicev GPT-4 v načrtu peskovnika.
- **Self-hosting Dify Community Edition</br>**
Hitro zaženite Dify v svojem okolju s tem [začetnim vodnikom](#quick-start) . Za dodatne reference in podrobnejša navodila uporabite našo [dokumentacijo](https://docs.dify.ai) .
- **Dify za podjetja/organizacije</br>**
Ponujamo dodatne funkcije, osredotočene na podjetja. Zabeležite svoja vprašanja prek tega klepetalnega robota ali nam pošljite e-pošto, da se pogovorimo o potrebah podjetja. </br>
> Za novoustanovljena podjetja in mala podjetja, ki uporabljajo AWS, si oglejte Dify Premium na AWS Marketplace in ga z enim klikom uvedite v svoj AWS VPC. To je cenovno ugodna ponudba AMI z možnostjo ustvarjanja aplikacij z logotipom in blagovno znamko po meri.
## Staying ahead
Star Dify on GitHub and be instantly notified of new releases.
![star-us](https://github.com/langgenius/dify/assets/13230914/b823edc1-6388-4e25-ad45-2f6b187adbb4)
## Napredne nastavitve
Če morate prilagoditi konfiguracijo, si oglejte komentarje v naši datoteki .env.example in posodobite ustrezne vrednosti v svoji .env datoteki. Poleg tega boste morda morali prilagoditi docker-compose.yamlsamo datoteko, na primer spremeniti različice slike, preslikave vrat ali namestitve nosilca, glede na vaše specifično okolje in zahteve za uvajanje. Po kakršnih koli spremembah ponovno zaženite docker-compose up -d. Celoten seznam razpoložljivih spremenljivk okolja najdete tukaj .
Če želite konfigurirati visoko razpoložljivo nastavitev, so na voljo Helm Charts in datoteke YAML, ki jih prispeva skupnost, ki omogočajo uvedbo Difyja v Kubernetes.
- [Helm Chart by @LeoQuote](https://github.com/douban/charts/tree/master/charts/dify)
- [Helm Chart by @BorisPolonsky](https://github.com/BorisPolonsky/dify-helm)
- [YAML file by @Winson-030](https://github.com/Winson-030/dify-kubernetes)
- [YAML file by @wyy-holding](https://github.com/wyy-holding/dify-k8s)
#### Uporaba Terraform za uvajanje
namestite Dify v Cloud Platform z enim klikom z uporabo [terraform](https://www.terraform.io/)
##### Azure Global
- [Azure Terraform by @nikawang](https://github.com/nikawang/dify-azure-terraform)
##### Google Cloud
- [Google Cloud Terraform by @sotazum](https://github.com/DeNA/dify-google-cloud-terraform)
#### Uporaba AWS CDK za uvajanje
Uvedite Dify v AWS z uporabo [CDK](https://aws.amazon.com/cdk/)
##### AWS
- [AWS CDK by @KevinZhao](https://github.com/aws-samples/solution-for-deploying-dify-on-aws)
## Prispevam
Za tiste, ki bi radi prispevali kodo, si oglejte naš vodnik za prispevke . Hkrati vas prosimo, da podprete Dify tako, da ga delite na družbenih medijih ter na dogodkih in konferencah.
> Iščemo sodelavce za pomoč pri prevajanju Difyja v jezike, ki niso mandarinščina ali angleščina. Če želite pomagati, si oglejte i18n README za več informacij in nam pustite komentar v global-userskanalu našega strežnika skupnosti Discord .
## Skupnost in stik
* [Github Discussion](https://github.com/langgenius/dify/discussions). Najboljše za: izmenjavo povratnih informacij in postavljanje vprašanj.
* [GitHub Issues](https://github.com/langgenius/dify/issues). Najboljše za: hrošče, na katere naletite pri uporabi Dify.AI, in predloge funkcij. Oglejte si naš [vodnik za prispevke](https://github.com/langgenius/dify/blob/main/CONTRIBUTING.md).
* [Discord](https://discord.gg/FngNHpbcY7). Najboljše za: deljenje vaših aplikacij in druženje s skupnostjo.
* [X(Twitter)](https://twitter.com/dify_ai). Najboljše za: deljenje vaših aplikacij in druženje s skupnostjo.
**Contributors**
<a href="https://github.com/langgenius/dify/graphs/contributors">
<img src="https://contrib.rocks/image?repo=langgenius/dify" />
</a>
## Star history
[![Star History Chart](https://api.star-history.com/svg?repos=langgenius/dify&type=Date)](https://star-history.com/#langgenius/dify&Date)
## Varnostno razkritje
Zaradi zaščite vaše zasebnosti se izogibajte objavljanju varnostnih vprašanj na GitHub. Namesto tega pošljite vprašanja na security@dify.ai in zagotovili vam bomo podrobnejši odgovor.
## Licenca
To skladišče je na voljo pod [odprtokodno licenco Dify](LICENSE) , ki je v bistvu Apache 2.0 z nekaj dodatnimi omejitvami.
![cover-v5-optimized](https://github.com/langgenius/dify/assets/13230914/f9e19af5-61ba-4119-b926-d10c4c06ebab)
<p align="center">
📌 <a href="https://dify.ai/blog/introducing-dify-workflow-file-upload-a-demo-on-ai-podcast">Predstavljamo nalaganje datotek Dify Workflow: znova ustvarite Google NotebookLM Podcast</a>
</p>
<p align="center">
<a href="https://cloud.dify.ai">Dify Cloud</a> ·
<a href="https://docs.dify.ai/getting-started/install-self-hosted">Samostojno gostovanje</a> ·
<a href="https://docs.dify.ai">Dokumentacija</a> ·
<a href="https://dify.ai/pricing">Pregled ponudb izdelkov Dify</a>
</p>
<p align="center">
<a href="https://dify.ai" target="_blank">
<img alt="Static Badge" src="https://img.shields.io/badge/Product-F04438"></a>
<a href="https://dify.ai/pricing" target="_blank">
<img alt="Static Badge" src="https://img.shields.io/badge/free-pricing?logo=free&color=%20%23155EEF&label=pricing&labelColor=%20%23528bff"></a>
<a href="https://discord.gg/FngNHpbcY7" target="_blank">
<img src="https://img.shields.io/discord/1082486657678311454?logo=discord&labelColor=%20%235462eb&logoColor=%20%23f5f5f5&color=%20%235462eb"
alt="chat on Discord"></a>
<a href="https://twitter.com/intent/follow?screen_name=dify_ai" target="_blank">
<img src="https://img.shields.io/twitter/follow/dify_ai?logo=X&color=%20%23f5f5f5"
alt="follow on X(Twitter)"></a>
<a href="https://www.linkedin.com/company/langgenius/" target="_blank">
<img src="https://custom-icon-badges.demolab.com/badge/LinkedIn-0A66C2?logo=linkedin-white&logoColor=fff"
alt="follow on LinkedIn"></a>
<a href="https://hub.docker.com/u/langgenius" target="_blank">
<img alt="Docker Pulls" src="https://img.shields.io/docker/pulls/langgenius/dify-web?labelColor=%20%23FDB062&color=%20%23f79009"></a>
<a href="https://github.com/langgenius/dify/graphs/commit-activity" target="_blank">
<img alt="Commits last month" src="https://img.shields.io/github/commit-activity/m/langgenius/dify?labelColor=%20%2332b583&color=%20%2312b76a"></a>
<a href="https://github.com/langgenius/dify/" target="_blank">
<img alt="Issues closed" src="https://img.shields.io/github/issues-search?query=repo%3Alanggenius%2Fdify%20is%3Aclosed&label=issues%20closed&labelColor=%20%237d89b0&color=%20%235d6b98"></a>
<a href="https://github.com/langgenius/dify/discussions/" target="_blank">
<img alt="Discussion posts" src="https://img.shields.io/github/discussions/langgenius/dify?labelColor=%20%239b8afb&color=%20%237a5af8"></a>
</p>
<p align="center">
<a href="./README.md"><img alt="README in English" src="https://img.shields.io/badge/English-d9d9d9"></a>
<a href="./README_CN.md"><img alt="简体中文版自述文件" src="https://img.shields.io/badge/简体中文-d9d9d9"></a>
<a href="./README_JA.md"><img alt="日本語のREADME" src="https://img.shields.io/badge/日本語-d9d9d9"></a>
<a href="./README_ES.md"><img alt="README en Español" src="https://img.shields.io/badge/Español-d9d9d9"></a>
<a href="./README_FR.md"><img alt="README en Français" src="https://img.shields.io/badge/Français-d9d9d9"></a>
<a href="./README_KL.md"><img alt="README tlhIngan Hol" src="https://img.shields.io/badge/Klingon-d9d9d9"></a>
<a href="./README_KR.md"><img alt="README in Korean" src="https://img.shields.io/badge/한국어-d9d9d9"></a>
<a href="./README_AR.md"><img alt="README بالعربية" src="https://img.shields.io/badge/العربية-d9d9d9"></a>
<a href="./README_TR.md"><img alt="Türkçe README" src="https://img.shields.io/badge/Türkçe-d9d9d9"></a>
<a href="./README_VI.md"><img alt="README Tiếng Việt" src="https://img.shields.io/badge/Ti%E1%BA%BFng%20Vi%E1%BB%87t-d9d9d9"></a>
<a href="./README_SI.md"><img alt="README Slovenščina" src="https://img.shields.io/badge/Sloven%C5%A1%C4%8Dina-d9d9d9"></a>
<a href="./README_BN.md"><img alt="README in বাংলা" src="https://img.shields.io/badge/বাংলা-d9d9d9"></a>
</p>
Dify je odprtokodna platforma za razvoj aplikacij LLM. Njegov intuitivni vmesnik združuje agentski potek dela z umetno inteligenco, cevovod RAG, zmogljivosti agentov, upravljanje modelov, funkcije opazovanja in več, kar vam omogoča hiter prehod od prototipa do proizvodnje.
## Hitri začetek
> Preden namestite Dify, se prepričajte, da vaša naprava izpolnjuje naslednje minimalne sistemske zahteve:
>
>- CPU >= 2 Core
>- RAM >= 4 GiB
</br>
Najlažji način za zagon strežnika Dify je prek docker compose . Preden zaženete Dify z naslednjimi ukazi, se prepričajte, da sta Docker in Docker Compose nameščena na vašem računalniku:
```bash
cd dify
cd docker
cp .env.example .env
docker compose up -d
```
Po zagonu lahko dostopate do nadzorne plošče Dify v brskalniku na [http://localhost/install](http://localhost/install) in začnete postopek inicializacije.
#### Iskanje pomoči
Prosimo, glejte naša pogosta vprašanja [FAQ](https://docs.dify.ai/getting-started/install-self-hosted/faqs) če naletite na težave pri nastavitvi Dify. Če imate še vedno težave, se obrnite na [skupnost ali nas](#community--contact).
> Če želite prispevati k Difyju ali narediti dodaten razvoj, glejte naš vodnik za [uvajanje iz izvorne kode](https://docs.dify.ai/getting-started/install-self-hosted/local-source-code)
## Ključne značilnosti
**1. Potek dela**:
Zgradite in preizkusite zmogljive poteke dela AI na vizualnem platnu, pri čemer izkoristite vse naslednje funkcije in več.
https://github.com/langgenius/dify/assets/13230914/356df23e-1604-483d-80a6-9517ece318aa
**2. Celovita podpora za modele**:
Brezhibna integracija s stotinami lastniških/odprtokodnih LLM-jev ducatov ponudnikov sklepanja in samostojnih rešitev, ki pokrivajo GPT, Mistral, Llama3 in vse modele, združljive z API-jem OpenAI. Celoten seznam podprtih ponudnikov modelov najdete [tukaj](https://docs.dify.ai/getting-started/readme/model-providers).
![providers-v5](https://github.com/langgenius/dify/assets/13230914/5a17bdbe-097a-4100-8363-40255b70f6e3)
**3. Prompt IDE**:
intuitivni vmesnik za ustvarjanje pozivov, primerjavo zmogljivosti modela in dodajanje dodatnih funkcij, kot je pretvorba besedila v govor, aplikaciji, ki temelji na klepetu.
**4. RAG Pipeline**:
E Obsežne zmogljivosti RAG, ki pokrivajo vse od vnosa dokumenta do priklica, s podporo za ekstrakcijo besedila iz datotek PDF, PPT in drugih običajnih formatov dokumentov.
**5. Agent capabilities**:
definirate lahko agente, ki temeljijo na klicanju funkcij LLM ali ReAct, in dodate vnaprej izdelana orodja ali orodja po meri za agenta. Dify ponuja več kot 50 vgrajenih orodij za agente AI, kot so Google Search, DALL·E, Stable Diffusion in WolframAlpha.
**6. LLMOps**:
Spremljajte in analizirajte dnevnike aplikacij in učinkovitost skozi čas. Pozive, nabore podatkov in modele lahko nenehno izboljšujete na podlagi proizvodnih podatkov in opomb.
**7. Backend-as-a-Service**:
AVse ponudbe Difyja so opremljene z ustreznimi API-ji, tako da lahko Dify brez težav integrirate v svojo poslovno logiko.
## Primerjava Funkcij
<table style="width: 100%;">
<tr>
<th align="center">Funkcija</th>
<th align="center">Dify.AI</th>
<th align="center">LangChain</th>
<th align="center">Flowise</th>
<th align="center">OpenAI Assistants API</th>
</tr>
<tr>
<td align="center">Programski pristop</td>
<td align="center">API + usmerjeno v aplikacije</td>
<td align="center">Python koda</td>
<td align="center">Usmerjeno v aplikacije</td>
<td align="center">Usmerjeno v API</td>
</tr>
<tr>
<td align="center">Podprti LLM-ji</td>
<td align="center">Bogata izbira</td>
<td align="center">Bogata izbira</td>
<td align="center">Bogata izbira</td>
<td align="center">Samo OpenAI</td>
</tr>
<tr>
<td align="center">RAG pogon</td>
<td align="center">✅</td>
<td align="center">✅</td>
<td align="center">✅</td>
<td align="center">✅</td>
</tr>
<tr>
<td align="center">Agent</td>
<td align="center">✅</td>
<td align="center">✅</td>
<td align="center">❌</td>
<td align="center">✅</td>
</tr>
<tr>
<td align="center">Potek dela</td>
<td align="center">✅</td>
<td align="center">❌</td>
<td align="center">✅</td>
<td align="center">❌</td>
</tr>
<tr>
<td align="center">Spremljanje</td>
<td align="center">✅</td>
<td align="center">✅</td>
<td align="center">❌</td>
<td align="center">❌</td>
</tr>
<tr>
<td align="center">Funkcija za podjetja (SSO/nadzor dostopa)</td>
<td align="center">✅</td>
<td align="center">❌</td>
<td align="center">❌</td>
<td align="center">❌</td>
</tr>
<tr>
<td align="center">Lokalna namestitev</td>
<td align="center">✅</td>
<td align="center">✅</td>
<td align="center">✅</td>
<td align="center">❌</td>
</tr>
</table>
## Uporaba Dify
- **Cloud </br>**
Gostimo storitev Dify Cloud za vsakogar, ki jo lahko preizkusite brez nastavitev. Zagotavlja vse zmožnosti različice za samostojno namestitev in vključuje 200 brezplačnih klicev GPT-4 v načrtu peskovnika.
- **Self-hosting Dify Community Edition</br>**
Hitro zaženite Dify v svojem okolju s tem [začetnim vodnikom](#quick-start) . Za dodatne reference in podrobnejša navodila uporabite našo [dokumentacijo](https://docs.dify.ai) .
- **Dify za podjetja/organizacije</br>**
Ponujamo dodatne funkcije, osredotočene na podjetja. Zabeležite svoja vprašanja prek tega klepetalnega robota ali nam pošljite e-pošto, da se pogovorimo o potrebah podjetja. </br>
> Za novoustanovljena podjetja in mala podjetja, ki uporabljajo AWS, si oglejte Dify Premium na AWS Marketplace in ga z enim klikom uvedite v svoj AWS VPC. To je cenovno ugodna ponudba AMI z možnostjo ustvarjanja aplikacij z logotipom in blagovno znamko po meri.
## Staying ahead
Star Dify on GitHub and be instantly notified of new releases.
![star-us](https://github.com/langgenius/dify/assets/13230914/b823edc1-6388-4e25-ad45-2f6b187adbb4)
## Napredne nastavitve
Če morate prilagoditi konfiguracijo, si oglejte komentarje v naši datoteki .env.example in posodobite ustrezne vrednosti v svoji .env datoteki. Poleg tega boste morda morali prilagoditi docker-compose.yamlsamo datoteko, na primer spremeniti različice slike, preslikave vrat ali namestitve nosilca, glede na vaše specifično okolje in zahteve za uvajanje. Po kakršnih koli spremembah ponovno zaženite docker-compose up -d. Celoten seznam razpoložljivih spremenljivk okolja najdete tukaj .
Če želite konfigurirati visoko razpoložljivo nastavitev, so na voljo Helm Charts in datoteke YAML, ki jih prispeva skupnost, ki omogočajo uvedbo Difyja v Kubernetes.
- [Helm Chart by @LeoQuote](https://github.com/douban/charts/tree/master/charts/dify)
- [Helm Chart by @BorisPolonsky](https://github.com/BorisPolonsky/dify-helm)
- [YAML file by @Winson-030](https://github.com/Winson-030/dify-kubernetes)
- [YAML file by @wyy-holding](https://github.com/wyy-holding/dify-k8s)
#### Uporaba Terraform za uvajanje
namestite Dify v Cloud Platform z enim klikom z uporabo [terraform](https://www.terraform.io/)
##### Azure Global
- [Azure Terraform by @nikawang](https://github.com/nikawang/dify-azure-terraform)
##### Google Cloud
- [Google Cloud Terraform by @sotazum](https://github.com/DeNA/dify-google-cloud-terraform)
#### Uporaba AWS CDK za uvajanje
Uvedite Dify v AWS z uporabo [CDK](https://aws.amazon.com/cdk/)
##### AWS
- [AWS CDK by @KevinZhao](https://github.com/aws-samples/solution-for-deploying-dify-on-aws)
## Prispevam
Za tiste, ki bi radi prispevali kodo, si oglejte naš vodnik za prispevke . Hkrati vas prosimo, da podprete Dify tako, da ga delite na družbenih medijih ter na dogodkih in konferencah.
> Iščemo sodelavce za pomoč pri prevajanju Difyja v jezike, ki niso mandarinščina ali angleščina. Če želite pomagati, si oglejte i18n README za več informacij in nam pustite komentar v global-userskanalu našega strežnika skupnosti Discord .
## Skupnost in stik
* [Github Discussion](https://github.com/langgenius/dify/discussions). Najboljše za: izmenjavo povratnih informacij in postavljanje vprašanj.
* [GitHub Issues](https://github.com/langgenius/dify/issues). Najboljše za: hrošče, na katere naletite pri uporabi Dify.AI, in predloge funkcij. Oglejte si naš [vodnik za prispevke](https://github.com/langgenius/dify/blob/main/CONTRIBUTING.md).
* [Discord](https://discord.gg/FngNHpbcY7). Najboljše za: deljenje vaših aplikacij in druženje s skupnostjo.
* [X(Twitter)](https://twitter.com/dify_ai). Najboljše za: deljenje vaših aplikacij in druženje s skupnostjo.
**Contributors**
<a href="https://github.com/langgenius/dify/graphs/contributors">
<img src="https://contrib.rocks/image?repo=langgenius/dify" />
</a>
## Star history
[![Star History Chart](https://api.star-history.com/svg?repos=langgenius/dify&type=Date)](https://star-history.com/#langgenius/dify&Date)
## Varnostno razkritje
Zaradi zaščite vaše zasebnosti se izogibajte objavljanju varnostnih vprašanj na GitHub. Namesto tega pošljite vprašanja na security@dify.ai in zagotovili vam bomo podrobnejši odgovor.
## Licenca
To skladišče je na voljo pod [odprtokodno licenco Dify](LICENSE) , ki je v bistvu Apache 2.0 z nekaj dodatnimi omejitvami.

2
api/.dockerignore
Unescape Escape View File

@ -16,4 +16,4 @@ logs
.ruff_cache
# venv
.venv
.venv

2
api/app_factory.py
Unescape Escape View File

@ -52,7 +52,6 @@ def initialize_extensions(app: DifyApp):
ext_mail,
ext_migrate,
ext_otel,
ext_otel_patch,
ext_proxy_fix,
ext_redis,
ext_repositories,
@ -85,7 +84,6 @@ def initialize_extensions(app: DifyApp):
ext_proxy_fix,
ext_blueprints,
ext_commands,
ext_otel_patch, # Apply patch before initializing OpenTelemetry
ext_otel,
]
for ext in extensions:

79
api/commands.py
Unescape Escape View File

@ -444,13 +444,13 @@ def convert_to_agent_apps():
WHERE a.mode = 'chat'
AND am.agent_mode is not null
AND (
am.agent_mode like '%"strategy": "function_call"%'
am.agent_mode like '%"strategy": "function_call"%'
OR am.agent_mode like '%"strategy": "react"%'
)
)
AND (
am.agent_mode like '{"enabled": true%'
am.agent_mode like '{"enabled": true%'
OR am.agent_mode like '{"max_iteration": %'
) ORDER BY a.created_at DESC LIMIT 1000
) ORDER BY a.created_at DESC LIMIT 1000
"""
with db.engine.begin() as conn:
@ -818,8 +818,9 @@ def clear_free_plan_tenant_expired_logs(days: int, batch: int, tenant_ids: list[
click.echo(click.style("Clear free plan tenant expired logs completed.", fg="green"))
@click.option("-f", "--force", is_flag=True, help="Skip user confirmation and force the command to execute.")
@click.command("clear-orphaned-file-records", help="Clear orphaned file records.")
def clear_orphaned_file_records():
def clear_orphaned_file_records(force: bool):
"""
Clear orphaned file records in the database.
"""
@ -845,7 +846,15 @@ def clear_orphaned_file_records():
# notify user and ask for confirmation
click.echo(
click.style("This command will find and delete orphaned file records in the following tables:", fg="yellow")
click.style(
"This command will first find and delete orphaned file records from the message_files table,", fg="yellow"
)
)
click.echo(
click.style(
"and then it will find and delete orphaned file records in the following tables:",
fg="yellow",
)
)
for files_table in files_tables:
click.echo(click.style(f"- {files_table['table']}", fg="yellow"))
@ -878,11 +887,55 @@ def clear_orphaned_file_records():
fg="yellow",
)
)
click.confirm("Do you want to proceed?", abort=True)
if not force:
click.confirm("Do you want to proceed?", abort=True)
# start the cleanup process
click.echo(click.style("Starting orphaned file records cleanup.", fg="white"))
# clean up the orphaned records in the message_files table where message_id doesn't exist in messages table
try:
click.echo(
click.style("- Listing message_files records where message_id doesn't exist in messages table", fg="white")
)
query = (
"SELECT mf.id, mf.message_id "
"FROM message_files mf LEFT JOIN messages m ON mf.message_id = m.id "
"WHERE m.id IS NULL"
)
orphaned_message_files = []
with db.engine.begin() as conn:
rs = conn.execute(db.text(query))
for i in rs:
orphaned_message_files.append({"id": str(i[0]), "message_id": str(i[1])})
if orphaned_message_files:
click.echo(click.style(f"Found {len(orphaned_message_files)} orphaned message_files records:", fg="white"))
for record in orphaned_message_files:
click.echo(click.style(f" - id: {record['id']}, message_id: {record['message_id']}", fg="black"))
if not force:
click.confirm(
(
f"Do you want to proceed "
f"to delete all {len(orphaned_message_files)} orphaned message_files records?"
),
abort=True,
)
click.echo(click.style("- Deleting orphaned message_files records", fg="white"))
query = "DELETE FROM message_files WHERE id IN :ids"
with db.engine.begin() as conn:
conn.execute(db.text(query), {"ids": tuple([record["id"] for record in orphaned_message_files])})
click.echo(
click.style(f"Removed {len(orphaned_message_files)} orphaned message_files records.", fg="green")
)
else:
click.echo(click.style("No orphaned message_files records found. There is nothing to delete.", fg="green"))
except Exception as e:
click.echo(click.style(f"Error deleting orphaned message_files records: {str(e)}", fg="red"))
# clean up the orphaned records in the rest of the *_files tables
try:
# fetch file id and keys from each table
all_files_in_tables = []
@ -964,7 +1017,8 @@ def clear_orphaned_file_records():
click.echo(click.style(f"Found {len(orphaned_files)} orphaned file records.", fg="white"))
for file in orphaned_files:
click.echo(click.style(f"- orphaned file id: {file}", fg="black"))
click.confirm(f"Do you want to proceed to delete all {len(orphaned_files)} orphaned file records?", abort=True)
if not force:
click.confirm(f"Do you want to proceed to delete all {len(orphaned_files)} orphaned file records?", abort=True)
# delete orphaned records for each file
try:
@ -979,8 +1033,9 @@ def clear_orphaned_file_records():
click.echo(click.style(f"Removed {len(orphaned_files)} orphaned file records.", fg="green"))
@click.option("-f", "--force", is_flag=True, help="Skip user confirmation and force the command to execute.")
@click.command("remove-orphaned-files-on-storage", help="Remove orphaned files on the storage.")
def remove_orphaned_files_on_storage():
def remove_orphaned_files_on_storage(force: bool):
"""
Remove orphaned files on the storage.
"""
@ -1028,7 +1083,8 @@ def remove_orphaned_files_on_storage():
fg="yellow",
)
)
click.confirm("Do you want to proceed?", abort=True)
if not force:
click.confirm("Do you want to proceed?", abort=True)
# start the cleanup process
click.echo(click.style("Starting orphaned files cleanup.", fg="white"))
@ -1069,7 +1125,8 @@ def remove_orphaned_files_on_storage():
click.echo(click.style(f"Found {len(orphaned_files)} orphaned files.", fg="white"))
for file in orphaned_files:
click.echo(click.style(f"- orphaned file: {file}", fg="black"))
click.confirm(f"Do you want to proceed to remove all {len(orphaned_files)} orphaned files?", abort=True)
if not force:
click.confirm(f"Do you want to proceed to remove all {len(orphaned_files)} orphaned files?", abort=True)
# delete orphaned files
removed_files = 0

31
api/configs/middleware/vdb/opensearch_config.py
Unescape Escape View File

@ -1,4 +1,5 @@
from typing import Optional
import enum
from typing import Literal, Optional
from pydantic import Field, PositiveInt
from pydantic_settings import BaseSettings
@ -9,6 +10,14 @@ class OpenSearchConfig(BaseSettings):
Configuration settings for OpenSearch
"""
class AuthMethod(enum.StrEnum):
"""
Authentication method for OpenSearch
"""
BASIC = "basic"
AWS_MANAGED_IAM = "aws_managed_iam"
OPENSEARCH_HOST: Optional[str] = Field(
description="Hostname or IP address of the OpenSearch server (e.g., 'localhost' or 'opensearch.example.com')",
default=None,
@ -19,6 +28,16 @@ class OpenSearchConfig(BaseSettings):
default=9200,
)
OPENSEARCH_SECURE: bool = Field(
description="Whether to use SSL/TLS encrypted connection for OpenSearch (True for HTTPS, False for HTTP)",
default=False,
)
OPENSEARCH_AUTH_METHOD: AuthMethod = Field(
description="Authentication method for OpenSearch connection (default is 'basic')",
default=AuthMethod.BASIC,
)
OPENSEARCH_USER: Optional[str] = Field(
description="Username for authenticating with OpenSearch",
default=None,
@ -29,7 +48,11 @@ class OpenSearchConfig(BaseSettings):
default=None,
)
OPENSEARCH_SECURE: bool = Field(
description="Whether to use SSL/TLS encrypted connection for OpenSearch (True for HTTPS, False for HTTP)",
default=False,
OPENSEARCH_AWS_REGION: Optional[str] = Field(
description="AWS region for OpenSearch (e.g. 'us-west-2')",
default=None,
)
OPENSEARCH_AWS_SERVICE: Optional[Literal["es", "aoss"]] = Field(
description="AWS service for OpenSearch (e.g. 'aoss' for OpenSearch Serverless)", default=None
)

2
api/configs/packaging/__init__.py
Unescape Escape View File

@ -9,7 +9,7 @@ class PackagingInfo(BaseSettings):
CURRENT_VERSION: str = Field(
description="Dify version",
default="1.3.0",
default="1.3.1",
)
COMMIT_SHA: str = Field(

18
api/constants/__init__.py
Unescape Escape View File

@ -16,11 +16,25 @@ AUDIO_EXTENSIONS.extend([ext.upper() for ext in AUDIO_EXTENSIONS])
if dify_config.ETL_TYPE == "Unstructured":
DOCUMENT_EXTENSIONS = ["txt", "markdown", "md", "mdx", "pdf", "html", "htm", "xlsx", "xls"]
DOCUMENT_EXTENSIONS = ["txt", "markdown", "md", "mdx", "pdf", "html", "htm", "xlsx", "xls", "vtt", "properties"]
DOCUMENT_EXTENSIONS.extend(("doc", "docx", "csv", "eml", "msg", "pptx", "xml", "epub"))
if dify_config.UNSTRUCTURED_API_URL:
DOCUMENT_EXTENSIONS.append("ppt")
DOCUMENT_EXTENSIONS.extend([ext.upper() for ext in DOCUMENT_EXTENSIONS])
else:
DOCUMENT_EXTENSIONS = ["txt", "markdown", "md", "mdx", "pdf", "html", "htm", "xlsx", "xls", "docx", "csv"]
DOCUMENT_EXTENSIONS = [
"txt",
"markdown",
"md",
"mdx",
"pdf",
"html",
"htm",
"xlsx",
"xls",
"docx",
"csv",
"vtt",
"properties",
]
DOCUMENT_EXTENSIONS.extend([ext.upper() for ext in DOCUMENT_EXTENSIONS])

2
api/controllers/files/image_preview.py
Unescape Escape View File

@ -89,7 +89,7 @@ class FilePreviewApi(Resource):
if args["as_attachment"]:
encoded_filename = quote(upload_file.name)
response.headers["Content-Disposition"] = f"attachment; filename*=UTF-8''{encoded_filename}"
response.headers["Content-Type"] = "application/octet-stream"
response.headers["Content-Type"] = "application/octet-stream"
return response

4
api/core/agent/prompt/template.py
Unescape Escape View File

@ -1,4 +1,4 @@
ENGLISH_REACT_COMPLETION_PROMPT_TEMPLATES = """Respond to the human as helpfully and accurately as possible.
ENGLISH_REACT_COMPLETION_PROMPT_TEMPLATES = """Respond to the human as helpfully and accurately as possible.
{{instruction}}
@ -47,7 +47,7 @@ Thought:""" # noqa: E501
ENGLISH_REACT_COMPLETION_AGENT_SCRATCHPAD_TEMPLATES = """Observation: {{observation}}
Thought:"""
ENGLISH_REACT_CHAT_PROMPT_TEMPLATES = """Respond to the human as helpfully and accurately as possible.
ENGLISH_REACT_CHAT_PROMPT_TEMPLATES = """Respond to the human as helpfully and accurately as possible.
{{instruction}}

2
api/core/app/task_pipeline/workflow_cycle_manage.py
Unescape Escape View File

@ -381,6 +381,8 @@ class WorkflowCycleManage:
workflow_node_execution.elapsed_time = elapsed_time
workflow_node_execution.execution_metadata = execution_metadata
self._workflow_node_execution_repository.update(workflow_node_execution)
return workflow_node_execution
def _handle_workflow_node_execution_retried(

2
api/core/external_data_tool/api/__builtin__
Unescape Escape View File

@ -1 +1 @@
1
1

6
api/core/helper/code_executor/javascript/javascript_transformer.py
Unescape Escape View File

@ -10,13 +10,13 @@ class NodeJsTemplateTransformer(TemplateTransformer):
f"""
// declare main function
{cls._code_placeholder}
// decode and prepare input object
var inputs_obj = JSON.parse(Buffer.from('{cls._inputs_placeholder}', 'base64').toString('utf-8'))
// execute main function
var output_obj = main(inputs_obj)
// convert output to json and print
var output_json = JSON.stringify(output_obj)
var result = `<<RESULT>>${{output_json}}<<RESULT>>`

16
api/core/helper/code_executor/jinja2/jinja2_transformer.py
Unescape Escape View File

@ -21,20 +21,20 @@ class Jinja2TemplateTransformer(TemplateTransformer):
import jinja2
template = jinja2.Template('''{cls._code_placeholder}''')
return template.render(**inputs)
import json
from base64 import b64decode
# decode and prepare input dict
inputs_obj = json.loads(b64decode('{cls._inputs_placeholder}').decode('utf-8'))
# execute main function
output = main(**inputs_obj)
# convert output and print
result = f'''<<RESULT>>{{output}}<<RESULT>>'''
print(result)
""")
return runner_script
@ -43,15 +43,15 @@ class Jinja2TemplateTransformer(TemplateTransformer):
preload_script = dedent("""
import jinja2
from base64 import b64decode
def _jinja2_preload_():
# prepare jinja2 environment, load template and render before to avoid sandbox issue
template = jinja2.Template('{{s}}')
template.render(s='a')
if __name__ == '__main__':
_jinja2_preload_()
""")
return preload_script

8
api/core/helper/code_executor/python3/python3_transformer.py
Unescape Escape View File

@ -9,16 +9,16 @@ class Python3TemplateTransformer(TemplateTransformer):
runner_script = dedent(f"""
# declare main function
{cls._code_placeholder}
import json
from base64 import b64decode
# decode and prepare input dict
inputs_obj = json.loads(b64decode('{cls._inputs_placeholder}').decode('utf-8'))
# execute main function
output_obj = main(**inputs_obj)
# convert output to json and print
output_json = json.dumps(output_obj, indent=4)
result = f'''<<RESULT>>{{output_json}}<<RESULT>>'''

17
api/core/llm_generator/llm_generator.py
Unescape Escape View File

@ -3,6 +3,8 @@ import logging
import re
from typing import Optional, cast
import json_repair
from core.llm_generator.output_parser.rule_config_generator import RuleConfigGeneratorOutputParser
from core.llm_generator.output_parser.suggested_questions_after_answer import SuggestedQuestionsAfterAnswerOutputParser
from core.llm_generator.prompts import (
@ -366,7 +368,20 @@ class LLMGenerator:
),
)
generated_json_schema = cast(str, response.message.content)
raw_content = response.message.content
if not isinstance(raw_content, str):
raise ValueError(f"LLM response content must be a string, got: {type(raw_content)}")
try:
parsed_content = json.loads(raw_content)
except json.JSONDecodeError:
parsed_content = json_repair.loads(raw_content)
if not isinstance(parsed_content, dict | list):
raise ValueError(f"Failed to parse structured output from llm: {raw_content}")
generated_json_schema = json.dumps(parsed_content, indent=2, ensure_ascii=False)
return {"output": generated_json_schema, "error": ""}
except InvokeError as e:

44
api/core/llm_generator/prompts.py
Unescape Escape View File

@ -1,5 +1,5 @@
# Written by YORKI MINAKO🤡, Edited by Xiaoyi
CONVERSATION_TITLE_PROMPT = """You need to decompose the user's input into "subject" and "intention" in order to accurately figure out what the user's input language actually is.
CONVERSATION_TITLE_PROMPT = """You need to decompose the user's input into "subject" and "intention" in order to accurately figure out what the user's input language actually is.
Notice: the language type user uses could be diverse, which can be English, Chinese, Italian, Español, Arabic, Japanese, French, and etc.
ENSURE your output is in the SAME language as the user's input!
Your output is restricted only to: (Input language) Intention + Subject(short as possible)
@ -58,7 +58,7 @@ User Input: yo, 你今天咋样？
"Your Output": "查询今日我的状态☺️"
}
User Input:
User Input:
""" # noqa: E501
PYTHON_CODE_GENERATOR_PROMPT_TEMPLATE = (
@ -163,11 +163,11 @@ Here is a task description for which I would like you to create a high-quality p
{{TASK_DESCRIPTION}}
</task_description>
Based on task description, please create a well-structured prompt template that another AI could use to consistently complete the task. The prompt template should include:
- Do not include <input> or <output> section and variables in the prompt, assume user will add them at their own will.
- Clear instructions for the AI that will be using this prompt, demarcated with <instruction> tags. The instructions should provide step-by-step directions on how to complete the task using the input variables. Also Specifies in the instructions that the output should not contain any xml tag.
- Relevant examples if needed to clarify the task further, demarcated with <example> tags. Do not include variables in the prompt. Give three pairs of input and output examples.
- Do not include <input> or <output> section and variables in the prompt, assume user will add them at their own will.
- Clear instructions for the AI that will be using this prompt, demarcated with <instruction> tags. The instructions should provide step-by-step directions on how to complete the task using the input variables. Also Specifies in the instructions that the output should not contain any xml tag.
- Relevant examples if needed to clarify the task further, demarcated with <example> tags. Do not include variables in the prompt. Give three pairs of input and output examples.
- Include other relevant sections demarcated with appropriate XML tags like <examples>, <instruction>.
- Use the same language as task description.
- Use the same language as task description.
- Output in ``` xml ``` and start with <instruction>
Please generate the full prompt template with at least 300 words and output only the prompt template.
""" # noqa: E501
@ -178,28 +178,28 @@ Here is a task description for which I would like you to create a high-quality p
{{TASK_DESCRIPTION}}
</task_description>
Based on task description, please create a well-structured prompt template that another AI could use to consistently complete the task. The prompt template should include:
- Descriptive variable names surrounded by {{ }} (two curly brackets) to indicate where the actual values will be substituted in. Choose variable names that clearly indicate the type of value expected. Variable names have to be composed of number, english alphabets and underline and nothing else.
- Clear instructions for the AI that will be using this prompt, demarcated with <instruction> tags. The instructions should provide step-by-step directions on how to complete the task using the input variables. Also Specifies in the instructions that the output should not contain any xml tag.
- Relevant examples if needed to clarify the task further, demarcated with <example> tags. Do not use curly brackets any other than in <instruction> section.
- Descriptive variable names surrounded by {{ }} (two curly brackets) to indicate where the actual values will be substituted in. Choose variable names that clearly indicate the type of value expected. Variable names have to be composed of number, english alphabets and underline and nothing else.
- Clear instructions for the AI that will be using this prompt, demarcated with <instruction> tags. The instructions should provide step-by-step directions on how to complete the task using the input variables. Also Specifies in the instructions that the output should not contain any xml tag.
- Relevant examples if needed to clarify the task further, demarcated with <example> tags. Do not use curly brackets any other than in <instruction> section.
- Any other relevant sections demarcated with appropriate XML tags like <input>, <output>, etc.
- Use the same language as task description.
- Use the same language as task description.
- Output in ``` xml ``` and start with <instruction>
Please generate the full prompt template and output only the prompt template.
""" # noqa: E501
RULE_CONFIG_PARAMETER_GENERATE_TEMPLATE = """
I need to extract the following information from the input text. The <information to be extracted> tag specifies the 'type', 'description' and 'required' of the information to be extracted.
I need to extract the following information from the input text. The <information to be extracted> tag specifies the 'type', 'description' and 'required' of the information to be extracted.
<information to be extracted>
variables name bounded two double curly brackets. Variable name has to be composed of number, english alphabets and underline and nothing else.
variables name bounded two double curly brackets. Variable name has to be composed of number, english alphabets and underline and nothing else.
</information to be extracted>
Step 1: Carefully read the input and understand the structure of the expected output.
Step 2: Extract relevant parameters from the provided text based on the name and description of object.
Step 2: Extract relevant parameters from the provided text based on the name and description of object.
Step 3: Structure the extracted parameters to JSON object as specified in <structure>.
Step 4: Ensure that the list of variable_names is properly formatted and valid. The output should not contain any XML tags. Output an empty list if there is no valid variable name in input text.
Step 4: Ensure that the list of variable_names is properly formatted and valid. The output should not contain any XML tags. Output an empty list if there is no valid variable name in input text.
### Structure
Here is the structure of the expected output, I should always follow the output structure.
Here is the structure of the expected output, I should always follow the output structure.
["variable_name_1", "variable_name_2"]
### Input Text
@ -214,13 +214,13 @@ I should always output a valid list. Output nothing other than the list of varia
RULE_CONFIG_STATEMENT_GENERATE_TEMPLATE = """
<instruction>
Step 1: Identify the purpose of the chatbot from the variable {{TASK_DESCRIPTION}} and infer chatbot's tone (e.g., friendly, professional, etc.) to add personality traits.
Step 1: Identify the purpose of the chatbot from the variable {{TASK_DESCRIPTION}} and infer chatbot's tone (e.g., friendly, professional, etc.) to add personality traits.
Step 2: Create a coherent and engaging opening statement.
Step 3: Ensure the output is welcoming and clearly explains what the chatbot is designed to do. Do not include any XML tags in the output.
Please use the same language as the user's input language. If user uses chinese then generate opening statement in chinese, if user uses english then generate opening statement in english.
Example Input:
Please use the same language as the user's input language. If user uses chinese then generate opening statement in chinese, if user uses english then generate opening statement in english.
Example Input:
Provide customer support for an e-commerce website
Example Output:
Example Output:
Welcome! I'm here to assist you with any questions or issues you might have with your shopping experience. Whether you're looking for product information, need help with your order, or have any other inquiries, feel free to ask. I'm friendly, helpful, and ready to support you in any way I can.
<Task>
Here is the task description: {{INPUT_TEXT}}
@ -276,15 +276,15 @@ Your task is to convert simple user descriptions into properly formatted JSON Sc
{
"type": "object",
"properties": {
"email": {
"email": {
"type": "string",
"format": "email"
},
"password": {
"password": {
"type": "string",
"minLength": 8
},
"age": {
"age": {
"type": "integer",
"minimum": 18
}

2
api/core/model_runtime/docs/en_US/customizable_model_scale_out.md
Unescape Escape View File

@ -307,4 +307,4 @@ Runtime Errors:
"""
```
For interface method details, see: [Interfaces](./interfaces.md). For specific implementations, refer to: [llm.py](https://github.com/langgenius/dify-runtime/blob/main/lib/model_providers/anthropic/llm/llm.py).
For interface method details, see: [Interfaces](./interfaces.md). For specific implementations, refer to: [llm.py](https://github.com/langgenius/dify-runtime/blob/main/lib/model_providers/anthropic/llm/llm.py).

2
api/core/model_runtime/docs/en_US/predefined_model_scale_out.md
Unescape Escape View File

@ -170,4 +170,4 @@ Runtime Errors:
"""
```
For interface method explanations, see: [Interfaces](./interfaces.md). For detailed implementation, refer to: [llm.py](https://github.com/langgenius/dify-runtime/blob/main/lib/model_providers/anthropic/llm/llm.py).
For interface method explanations, see: [Interfaces](./interfaces.md). For detailed implementation, refer to: [llm.py](https://github.com/langgenius/dify-runtime/blob/main/lib/model_providers/anthropic/llm/llm.py).

2
api/core/model_runtime/docs/zh_Hans/customizable_model_scale_out.md
Unescape Escape View File

@ -294,4 +294,4 @@ provider_credential_schema:
"""
```
接口方法说明见：[Interfaces](./interfaces.md)，具体实现可参考：[llm.py](https://github.com/langgenius/dify-runtime/blob/main/lib/model_providers/anthropic/llm/llm.py)。
接口方法说明见：[Interfaces](./interfaces.md)，具体实现可参考：[llm.py](https://github.com/langgenius/dify-runtime/blob/main/lib/model_providers/anthropic/llm/llm.py)。

2
api/core/model_runtime/docs/zh_Hans/predefined_model_scale_out.md
Unescape Escape View File

@ -169,4 +169,4 @@ pricing: # 价格信息
"""
```
接口方法说明见：[Interfaces](./interfaces.md)，具体实现可参考：[llm.py](https://github.com/langgenius/dify-runtime/blob/main/lib/model_providers/anthropic/llm/llm.py)。
接口方法说明见：[Interfaces](./interfaces.md)，具体实现可参考：[llm.py](https://github.com/langgenius/dify-runtime/blob/main/lib/model_providers/anthropic/llm/llm.py)。

2
api/core/moderation/api/__builtin__
Unescape Escape View File

@ -1 +1 @@
3
3

2
api/core/moderation/keywords/__builtin__
Unescape Escape View File

@ -1 +1 @@
2
2

2
api/core/moderation/openai_moderation/__builtin__
Unescape Escape View File

@ -1 +1 @@
1
1

2
api/core/plugin/backwards_invocation/app.py
Unescape Escape View File

@ -72,7 +72,7 @@ class PluginAppBackwardsInvocation(BaseBackwardsInvocation):
raise ValueError("missing query")
return cls.invoke_chat_app(app, user, conversation_id, query, stream, inputs, files)
elif app.mode == AppMode.WORKFLOW.value:
elif app.mode == AppMode.WORKFLOW:
return cls.invoke_workflow_app(app, user, stream, inputs, files)
elif app.mode == AppMode.COMPLETION:
return cls.invoke_completion_app(app, user, stream, inputs, files)

4
api/core/plugin/backwards_invocation/model.py
Unescape Escape View File

@ -239,8 +239,8 @@ class PluginModelBackwardsInvocation(BaseBackwardsInvocation):
content = payload.text
SUMMARY_PROMPT = """You are a professional language researcher, you are interested in the language
and you can quickly aimed at the main point of an webpage and reproduce it in your own words but
retain the original meaning and keep the key points.
and you can quickly aimed at the main point of an webpage and reproduce it in your own words but
retain the original meaning and keep the key points.
however, the text you got is too long, what you got is possible a part of the text.
Please summarize the text you got.

2
api/core/prompt/prompt_templates/baichuan_chat.json
Unescape Escape View File

@ -10,4 +10,4 @@
],
"query_prompt": "\n\n用户：{{#query#}}",
"stops": ["用户："]
}
}

2
api/core/prompt/prompt_templates/baichuan_completion.json
Unescape Escape View File

@ -6,4 +6,4 @@
],
"query_prompt": "{{#query#}}",
"stops": null
}
}

2
api/core/prompt/prompt_templates/common_completion.json
Unescape Escape View File

@ -6,4 +6,4 @@
],
"query_prompt": "{{#query#}}",
"stops": null
}
}

6
api/core/rag/datasource/vdb/analyticdb/analyticdb_vector_sql.py
Unescape Escape View File

@ -156,8 +156,8 @@ class AnalyticdbVectorBySql:
values = []
id_prefix = str(uuid.uuid4()) + "_"
sql = f"""
INSERT INTO {self.table_name}
(id, ref_doc_id, vector, page_content, metadata_, to_tsvector)
INSERT INTO {self.table_name}
(id, ref_doc_id, vector, page_content, metadata_, to_tsvector)
VALUES (%s, %s, %s, %s, %s, to_tsvector('zh_cn', %s));
"""
for i, doc in enumerate(documents):
@ -242,7 +242,7 @@ class AnalyticdbVectorBySql:
where_clause += f"AND metadata_->>'document_id' IN ({document_ids})"
with self._get_cursor() as cur:
cur.execute(
f"""SELECT id, vector, page_content, metadata_,
f"""SELECT id, vector, page_content, metadata_,
ts_rank(to_tsvector, to_tsquery_from_text(%s, 'zh_cn'), 32) AS score
FROM {self.table_name}
WHERE to_tsvector@@to_tsquery_from_text(%s, 'zh_cn') {where_clause}

20
api/core/rag/datasource/vdb/milvus/milvus_vector.py
Unescape Escape View File

@ -27,8 +27,8 @@ class MilvusConfig(BaseModel):
uri: str # Milvus server URI
token: Optional[str] = None # Optional token for authentication
user: str # Username for authentication
password: str # Password for authentication
user: Optional[str] = None # Username for authentication
password: Optional[str] = None # Password for authentication
batch_size: int = 100 # Batch size for operations
database: str = "default" # Database name
enable_hybrid_search: bool = False # Flag to enable hybrid search
@ -43,10 +43,11 @@ class MilvusConfig(BaseModel):
"""
if not values.get("uri"):
raise ValueError("config MILVUS_URI is required")
if not values.get("user"):
raise ValueError("config MILVUS_USER is required")
if not values.get("password"):
raise ValueError("config MILVUS_PASSWORD is required")
if not values.get("token"):
if not values.get("user"):
raise ValueError("config MILVUS_USER is required")
if not values.get("password"):
raise ValueError("config MILVUS_PASSWORD is required")
return values
def to_milvus_params(self):
@ -356,11 +357,14 @@ class MilvusVector(BaseVector):
)
redis_client.set(collection_exist_cache_key, 1, ex=3600)
def _init_client(self, config) -> MilvusClient:
def _init_client(self, config: MilvusConfig) -> MilvusClient:
"""
Initialize and return a Milvus client.
"""
client = MilvusClient(uri=config.uri, user=config.user, password=config.password, db_name=config.database)
if config.token:
client = MilvusClient(uri=config.uri, token=config.token, db_name=config.database)
else:
client = MilvusClient(uri=config.uri, user=config.user, password=config.password, db_name=config.database)
return client

2
api/core/rag/datasource/vdb/oceanbase/oceanbase_vector.py
Unescape Escape View File

@ -203,7 +203,7 @@ class OceanBaseVector(BaseVector):
full_sql = f"""SELECT metadata, text, MATCH (text) AGAINST (:query) AS score
FROM {self._collection_name}
WHERE MATCH (text) AGAINST (:query) > 0
WHERE MATCH (text) AGAINST (:query) > 0
{where_clause}
ORDER BY score DESC
LIMIT {top_k}"""

4
api/core/rag/datasource/vdb/opengauss/opengauss.py
Unescape Escape View File

@ -59,12 +59,12 @@ CREATE TABLE IF NOT EXISTS {table_name} (
"""
SQL_CREATE_INDEX_PQ = """
CREATE INDEX IF NOT EXISTS embedding_{table_name}_pq_idx ON {table_name}
CREATE INDEX IF NOT EXISTS embedding_{table_name}_pq_idx ON {table_name}
USING hnsw (embedding vector_cosine_ops) WITH (m = 16, ef_construction = 64, enable_pq=on, pq_m={pq_m});
"""
SQL_CREATE_INDEX = """
CREATE INDEX IF NOT EXISTS embedding_cosine_{table_name}_idx ON {table_name}
CREATE INDEX IF NOT EXISTS embedding_cosine_{table_name}_idx ON {table_name}
USING hnsw (embedding vector_cosine_ops) WITH (m = 16, ef_construction = 64);
"""

59
api/core/rag/datasource/vdb/opensearch/opensearch_vector.py
Unescape Escape View File

@ -1,10 +1,9 @@
import json
import logging
import ssl
from typing import Any, Optional
from typing import Any, Literal, Optional
from uuid import uuid4
from opensearchpy import OpenSearch, helpers
from opensearchpy import OpenSearch, Urllib3AWSV4SignerAuth, Urllib3HttpConnection, helpers
from opensearchpy.helpers import BulkIndexError
from pydantic import BaseModel, model_validator
@ -24,9 +23,12 @@ logger = logging.getLogger(__name__)
class OpenSearchConfig(BaseModel):
host: str
port: int
secure: bool = False
auth_method: Literal["basic", "aws_managed_iam"] = "basic"
user: Optional[str] = None
password: Optional[str] = None
secure: bool = False
aws_region: Optional[str] = None
aws_service: Optional[str] = None
@model_validator(mode="before")
@classmethod
@ -35,24 +37,40 @@ class OpenSearchConfig(BaseModel):
raise ValueError("config OPENSEARCH_HOST is required")
if not values.get("port"):
raise ValueError("config OPENSEARCH_PORT is required")
if values.get("auth_method") == "aws_managed_iam":
if not values.get("aws_region"):
raise ValueError("config OPENSEARCH_AWS_REGION is required for AWS_MANAGED_IAM auth method")
if not values.get("aws_service"):
raise ValueError("config OPENSEARCH_AWS_SERVICE is required for AWS_MANAGED_IAM auth method")
return values
def create_ssl_context(self) -> ssl.SSLContext:
ssl_context = ssl.create_default_context()
ssl_context.check_hostname = False
ssl_context.verify_mode = ssl.CERT_NONE # Disable Certificate Validation
return ssl_context
def create_aws_managed_iam_auth(self) -> Urllib3AWSV4SignerAuth:
import boto3 # type: ignore
return Urllib3AWSV4SignerAuth(
credentials=boto3.Session().get_credentials(),
region=self.aws_region,
service=self.aws_service, # type: ignore[arg-type]
)
def to_opensearch_params(self) -> dict[str, Any]:
params = {
"hosts": [{"host": self.host, "port": self.port}],
"use_ssl": self.secure,
"verify_certs": self.secure,
"connection_class": Urllib3HttpConnection,
"pool_maxsize": 20,
}
if self.user and self.password:
if self.auth_method == "basic":
logger.info("Using basic authentication for OpenSearch Vector DB")
params["http_auth"] = (self.user, self.password)
if self.secure:
params["ssl_context"] = self.create_ssl_context()
elif self.auth_method == "aws_managed_iam":
logger.info("Using AWS managed IAM role for OpenSearch Vector DB")
params["http_auth"] = self.create_aws_managed_iam_auth()
return params
@ -76,16 +94,23 @@ class OpenSearchVector(BaseVector):
action = {
"_op_type": "index",
"_index": self._collection_name.lower(),
"_id": uuid4().hex,
"_source": {
Field.CONTENT_KEY.value: documents[i].page_content,
Field.VECTOR.value: embeddings[i], # Make sure you pass an array here
Field.METADATA_KEY.value: documents[i].metadata,
},
}
# See https://github.com/langchain-ai/langchainjs/issues/4346#issuecomment-1935123377
if self._client_config.aws_service not in ["aoss"]:
action["_id"] = uuid4().hex
actions.append(action)
helpers.bulk(self._client, actions)
helpers.bulk(
client=self._client,
actions=actions,
timeout=30,
max_retries=3,
)
def get_ids_by_metadata_field(self, key: str, value: str):
query = {"query": {"term": {f"{Field.METADATA_KEY.value}.{key}": value}}}
@ -234,6 +259,7 @@ class OpenSearchVector(BaseVector):
},
}
logger.info(f"Creating OpenSearch index {self._collection_name.lower()}")
self._client.indices.create(index=self._collection_name.lower(), body=index_body)
redis_client.set(collection_exist_cache_key, 1, ex=3600)
@ -252,9 +278,12 @@ class OpenSearchVectorFactory(AbstractVectorFactory):
open_search_config = OpenSearchConfig(
host=dify_config.OPENSEARCH_HOST or "localhost",
port=dify_config.OPENSEARCH_PORT,
secure=dify_config.OPENSEARCH_SECURE,
auth_method=dify_config.OPENSEARCH_AUTH_METHOD.value,
user=dify_config.OPENSEARCH_USER,
password=dify_config.OPENSEARCH_PASSWORD,
secure=dify_config.OPENSEARCH_SECURE,
aws_region=dify_config.OPENSEARCH_AWS_REGION,
aws_service=dify_config.OPENSEARCH_AWS_SERVICE,
)
return OpenSearchVector(collection_name=collection_name, config=open_search_config)

12
api/core/rag/datasource/vdb/oracle/oraclevector.py
Unescape Escape View File

@ -59,8 +59,8 @@ CREATE TABLE IF NOT EXISTS {table_name} (
)
"""
SQL_CREATE_INDEX = """
CREATE INDEX IF NOT EXISTS idx_docs_{table_name} ON {table_name}(text)
INDEXTYPE IS CTXSYS.CONTEXT PARAMETERS
CREATE INDEX IF NOT EXISTS idx_docs_{table_name} ON {table_name}(text)
INDEXTYPE IS CTXSYS.CONTEXT PARAMETERS
('FILTER CTXSYS.NULL_FILTER SECTION GROUP CTXSYS.HTML_SECTION_GROUP LEXER world_lexer')
"""
@ -164,7 +164,7 @@ class OracleVector(BaseVector):
with conn.cursor() as cur:
try:
cur.execute(
f"""INSERT INTO {self.table_name} (id, text, meta, embedding)
f"""INSERT INTO {self.table_name} (id, text, meta, embedding)
VALUES (:1, :2, :3, :4)""",
value,
)
@ -227,8 +227,8 @@ class OracleVector(BaseVector):
conn.outputtypehandler = self.output_type_handler
with conn.cursor() as cur:
cur.execute(
f"""SELECT meta, text, vector_distance(embedding,(select to_vector(:1) from dual),cosine)
AS distance FROM {self.table_name}
f"""SELECT meta, text, vector_distance(embedding,(select to_vector(:1) from dual),cosine)
AS distance FROM {self.table_name}
{where_clause} ORDER BY distance fetch first {top_k} rows only""",
[numpy.array(query_vector)],
)
@ -290,7 +290,7 @@ class OracleVector(BaseVector):
document_ids = ", ".join(f"'{id}'" for id in document_ids_filter)
where_clause = f" AND metadata->>'document_id' in ({document_ids}) "
cur.execute(
f"""select meta, text, embedding FROM {self.table_name}
f"""select meta, text, embedding FROM {self.table_name}
WHERE CONTAINS(text, :kk, 1) > 0 {where_clause}
order by score(1) desc fetch first {top_k} rows only""",
kk=" ACCUM ".join(entities),

2
api/core/rag/datasource/vdb/pgvector/pgvector.py
Unescape Escape View File

@ -61,7 +61,7 @@ CREATE TABLE IF NOT EXISTS {table_name} (
"""
SQL_CREATE_INDEX = """
CREATE INDEX IF NOT EXISTS embedding_cosine_v1_idx ON {table_name}
CREATE INDEX IF NOT EXISTS embedding_cosine_v1_idx ON {table_name}
USING hnsw (embedding vector_cosine_ops) WITH (m = 16, ef_construction = 64);
"""

2
api/core/rag/datasource/vdb/pyvastbase/vastbase_vector.py
Unescape Escape View File

@ -58,7 +58,7 @@ CREATE TABLE IF NOT EXISTS {table_name} (
"""
SQL_CREATE_INDEX = """
CREATE INDEX IF NOT EXISTS embedding_cosine_v1_idx ON {table_name}
CREATE INDEX IF NOT EXISTS embedding_cosine_v1_idx ON {table_name}
USING hnsw (embedding floatvector_cosine_ops) WITH (m = 16, ef_construction = 64);
"""

4
api/core/rag/datasource/vdb/tidb_vector/tidb_vector.py
Unescape Escape View File

@ -205,9 +205,9 @@ class TiDBVector(BaseVector):
with Session(self._engine) as session:
select_statement = sql_text(f"""
SELECT meta, text, distance
SELECT meta, text, distance
FROM (
SELECT
SELECT
meta,
text,
{tidb_dist_func}(vector, :query_vector_str) AS distance

22
api/core/rag/rerank/rerank_model.py
Unescape Escape View File

@ -52,14 +52,16 @@ class RerankModelRunner(BaseRerankRunner):
rerank_documents = []
for result in rerank_result.docs:
# format document
rerank_document = Document(
page_content=result.text,
metadata=documents[result.index].metadata,
provider=documents[result.index].provider,
)
if rerank_document.metadata is not None:
rerank_document.metadata["score"] = result.score
rerank_documents.append(rerank_document)
if score_threshold is None or result.score >= score_threshold:
# format document
rerank_document = Document(
page_content=result.text,
metadata=documents[result.index].metadata,
provider=documents[result.index].provider,
)
if rerank_document.metadata is not None:
rerank_document.metadata["score"] = result.score
rerank_documents.append(rerank_document)
return rerank_documents
rerank_documents.sort(key=lambda x: x.metadata.get("score", 0.0), reverse=True)
return rerank_documents[:top_n] if top_n else rerank_documents

4
api/core/rag/retrieval/template_prompts.py
Unescape Escape View File

@ -50,7 +50,7 @@ You are a text metadata extract engine that extract text's metadata based on use
# Your task is to ONLY extract the metadatas that exist in the input text from the provided metadata list and Use the following operators ["=", "!=", ">", "<", ">=", "<="] to express logical relationships, then return result in JSON format with the key "metadata_fields" and value "metadata_field_value" and comparison operator "comparison_operator".
### Format
The input text is in the variable input_text. Metadata are specified as a list in the variable metadata_fields.
### Constraint
### Constraint
DO NOT include anything other than the JSON array in your response.
### Example
Here is the chat example between human and assistant, inside <example></example> XML tags.
@ -59,7 +59,7 @@ User:{{"input_text": ["I want to know which company’s email address test@examp
Assistant:{{"metadata_map": [{{"metadata_field_name": "email", "metadata_field_value": "test@example.com", "comparison_operator": "="}}]}}
User:{{"input_text": "What are the movies with a score of more than 9 in 2024?", "metadata_fields": ["name", "year", "rating", "country"]}}
Assistant:{{"metadata_map": [{{"metadata_field_name": "year", "metadata_field_value": "2024", "comparison_operator": "="}, {{"metadata_field_name": "rating", "metadata_field_value": "9", "comparison_operator": ">"}}]}}
</example>
</example>
### User Input
{{"input_text" : "{input_text}", "metadata_fields" : {metadata_fields}}}
### Assistant Output

4
api/core/tools/builtin_tool/tool.py
Unescape Escape View File

@ -6,8 +6,8 @@ from core.tools.entities.tool_entities import ToolProviderType
from core.tools.utils.model_invocation_utils import ModelInvocationUtils
_SUMMARY_PROMPT = """You are a professional language researcher, you are interested in the language
and you can quickly aimed at the main point of an webpage and reproduce it in your own words but
retain the original meaning and keep the key points.
and you can quickly aimed at the main point of an webpage and reproduce it in your own words but
retain the original meaning and keep the key points.
however, the text you got is too long, what you got is possible a part of the text.
Please summarize the text you got.
"""

2
api/core/tools/tool_engine.py
Unescape Escape View File

@ -246,7 +246,7 @@ class ToolEngine:
+ "you do not need to create it, just tell the user to check it now."
)
elif response.type == ToolInvokeMessage.MessageType.JSON:
result = json.dumps(
result += json.dumps(
cast(ToolInvokeMessage.JsonMessage, response.message).json_object, ensure_ascii=False
)
else:

82
api/core/workflow/nodes/document_extractor/node.py
Unescape Escape View File

@ -11,6 +11,7 @@ import docx
import pandas as pd
import pypandoc # type: ignore
import pypdfium2 # type: ignore
import webvtt # type: ignore
import yaml # type: ignore
from docx.document import Document
from docx.oxml.table import CT_Tbl
@ -132,6 +133,10 @@ def _extract_text_by_mime_type(*, file_content: bytes, mime_type: str) -> str:
return _extract_text_from_json(file_content)
case "application/x-yaml" | "text/yaml":
return _extract_text_from_yaml(file_content)
case "text/vtt":
return _extract_text_from_vtt(file_content)
case "text/properties":
return _extract_text_from_properties(file_content)
case _:
raise UnsupportedFileTypeError(f"Unsupported MIME type: {mime_type}")
@ -139,7 +144,7 @@ def _extract_text_by_mime_type(*, file_content: bytes, mime_type: str) -> str:
def _extract_text_by_file_extension(*, file_content: bytes, file_extension: str) -> str:
"""Extract text from a file based on its file extension."""
match file_extension:
case ".txt" | ".markdown" | ".md" | ".html" | ".htm" | ".xml" | ".vtt":
case ".txt" | ".markdown" | ".md" | ".html" | ".htm" | ".xml":
return _extract_text_from_plain_text(file_content)
case ".json":
return _extract_text_from_json(file_content)
@ -165,6 +170,10 @@ def _extract_text_by_file_extension(*, file_content: bytes, file_extension: str)
return _extract_text_from_eml(file_content)
case ".msg":
return _extract_text_from_msg(file_content)
case ".vtt":
return _extract_text_from_vtt(file_content)
case ".properties":
return _extract_text_from_properties(file_content)
case _:
raise UnsupportedFileTypeError(f"Unsupported Extension Type: {file_extension}")
@ -214,8 +223,8 @@ def _extract_text_from_doc(file_content: bytes) -> str:
"""
from unstructured.partition.api import partition_via_api
if not (dify_config.UNSTRUCTURED_API_URL and dify_config.UNSTRUCTURED_API_KEY):
raise TextExtractionError("UNSTRUCTURED_API_URL and UNSTRUCTURED_API_KEY must be set")
if not dify_config.UNSTRUCTURED_API_URL:
raise TextExtractionError("UNSTRUCTURED_API_URL must be set")
try:
with tempfile.NamedTemporaryFile(suffix=".doc", delete=False) as temp_file:
@ -226,7 +235,7 @@ def _extract_text_from_doc(file_content: bytes) -> str:
file=file,
metadata_filename=temp_file.name,
api_url=dify_config.UNSTRUCTURED_API_URL,
api_key=dify_config.UNSTRUCTURED_API_KEY,
api_key=dify_config.UNSTRUCTURED_API_KEY, # type: ignore
)
os.unlink(temp_file.name)
return "\n".join([getattr(element, "text", "") for element in elements])
@ -462,3 +471,68 @@ def _extract_text_from_msg(file_content: bytes) -> str:
return "\n".join([str(element) for element in elements])
except Exception as e:
raise TextExtractionError(f"Failed to extract text from MSG: {str(e)}") from e
def _extract_text_from_vtt(vtt_bytes: bytes) -> str:
text = _extract_text_from_plain_text(vtt_bytes)
# remove bom
text = text.lstrip("\ufeff")
raw_results = []
for caption in webvtt.from_string(text):
raw_results.append((caption.voice, caption.text))
# Merge consecutive utterances by the same speaker
merged_results = []
if raw_results:
current_speaker, current_text = raw_results[0]
for i in range(1, len(raw_results)):
spk, txt = raw_results[i]
if spk == None:
merged_results.append((None, current_text))
continue
if spk == current_speaker:
# If it is the same speaker, merge the utterances (joined by space)
current_text += " " + txt
else:
# If the speaker changes, register the utterance so far and move on
merged_results.append((current_speaker, current_text))
current_speaker, current_text = spk, txt
# Add the last element
merged_results.append((current_speaker, current_text))
else:
merged_results = raw_results
# Return the result in the specified format: Speaker "text" style
formatted = [f'{spk or ""} "{txt}"' for spk, txt in merged_results]
return "\n".join(formatted)
def _extract_text_from_properties(file_content: bytes) -> str:
try:
text = _extract_text_from_plain_text(file_content)
lines = text.splitlines()
result = []
for line in lines:
line = line.strip()
# Preserve comments and empty lines
if not line or line.startswith("#") or line.startswith("!"):
result.append(line)
continue
if "=" in line:
key, value = line.split("=", 1)
elif ":" in line:
key, value = line.split(":", 1)
else:
key, value = line, ""
result.append(f"{key.strip()}: {value.strip()}")
return "\n".join(result)
except Exception as e:
raise TextExtractionError(f"Failed to extract text from properties file: {str(e)}") from e

4
api/core/workflow/nodes/knowledge_retrieval/template_prompts.py
Unescape Escape View File

@ -50,7 +50,7 @@ You are a text metadata extract engine that extract text's metadata based on use
# Your task is to ONLY extract the metadatas that exist in the input text from the provided metadata list and Use the following operators ["=", "!=", ">", "<", ">=", "<="] to express logical relationships, then return result in JSON format with the key "metadata_fields" and value "metadata_field_value" and comparison operator "comparison_operator".
### Format
The input text is in the variable input_text. Metadata are specified as a list in the variable metadata_fields.
### Constraint
### Constraint
DO NOT include anything other than the JSON array in your response.
### Example
Here is the chat example between human and assistant, inside <example></example> XML tags.
@ -59,7 +59,7 @@ User:{{"input_text": ["I want to know which company’s email address test@examp
Assistant:{{"metadata_map": [{{"metadata_field_name": "email", "metadata_field_value": "test@example.com", "comparison_operator": "="}}]}}
User:{{"input_text": "What are the movies with a score of more than 9 in 2024?", "metadata_fields": ["name", "year", "rating", "country"]}}
Assistant:{{"metadata_map": [{{"metadata_field_name": "year", "metadata_field_value": "2024", "comparison_operator": "="}, {{"metadata_field_name": "rating", "metadata_field_value": "9", "comparison_operator": ">"}}]}}
</example>
</example>
### User Input
{{"input_text" : "{input_text}", "metadata_fields" : {metadata_fields}}}
### Assistant Output

14
api/core/workflow/nodes/parameter_extractor/prompts.py
Unescape Escape View File

@ -17,7 +17,7 @@ Some additional information is provided below. Always adhere to these instructio
</instruction>
Steps:
1. Review the chat history provided within the <histories> tags.
2. Extract the relevant information based on the criteria given, output multiple values if there is multiple relevant information that match the criteria in the given text.
2. Extract the relevant information based on the criteria given, output multiple values if there is multiple relevant information that match the criteria in the given text.
3. Generate a well-formatted output using the defined functions and arguments.
4. Use the `extract_parameter` function to create structured outputs with appropriate parameters.
5. Do not include any XML tags in your output.
@ -89,13 +89,13 @@ Some extra information are provided below, I should always follow the instructio
</instructions>
### Extract parameter Workflow
I need to extract the following information from the input text. The <information to be extracted> tag specifies the 'type', 'description' and 'required' of the information to be extracted.
I need to extract the following information from the input text. The <information to be extracted> tag specifies the 'type', 'description' and 'required' of the information to be extracted.
<information to be extracted>
{{ structure }}
</information to be extracted>
Step 1: Carefully read the input and understand the structure of the expected output.
Step 2: Extract relevant parameters from the provided text based on the name and description of object.
Step 2: Extract relevant parameters from the provided text based on the name and description of object.
Step 3: Structure the extracted parameters to JSON object as specified in <structure>.
Step 4: Ensure that the JSON object is properly formatted and valid. The output should not contain any XML tags. Only the JSON object should be outputted.
@ -106,10 +106,10 @@ Here are the chat histories between human and assistant, inside <histories></his
</histories>
### Structure
Here is the structure of the expected output, I should always follow the output structure.
Here is the structure of the expected output, I should always follow the output structure.
{{γγγ
'properties1': 'relevant text extracted from input',
'properties2': 'relevant text extracted from input',
'properties1': 'relevant text extracted from input',
'properties2': 'relevant text extracted from input',
}}γγγ
### Input Text
@ -119,7 +119,7 @@ Inside <text></text> XML tags, there is a text that I should extract parameters
</text>
### Answer
I should always output a valid JSON object. Output nothing other than the JSON object.
I should always output a valid JSON object. Output nothing other than the JSON object.
```JSON
""" # noqa: E501

4
api/core/workflow/nodes/question_classifier/template_prompts.py
Unescape Escape View File

@ -55,7 +55,7 @@ You are a text classification engine that analyzes text data and assigns categor
Your task is to assign one categories ONLY to the input text and only one category may be assigned returned in the output. Additionally, you need to extract the key words from the text that are related to the classification.
### Format
The input text is in the variable input_text. Categories are specified as a category list with two filed category_id and category_name in the variable categories. Classification instructions may be included to improve the classification accuracy.
### Constraint
### Constraint
DO NOT include anything other than the JSON array in your response.
### Example
Here is the chat example between human and assistant, inside <example></example> XML tags.
@ -64,7 +64,7 @@ User:{{"input_text": ["I recently had a great experience with your company. The
Assistant:{{"keywords": ["recently", "great experience", "company", "service", "prompt", "staff", "friendly"],"category_id": "f5660049-284f-41a7-b301-fd24176a711c","category_name": "Customer Service"}}
User:{{"input_text": ["bad service, slow to bring the food"], "categories": [{{"category_id":"80fb86a0-4454-4bf5-924c-f253fdd83c02","category_name":"Food Quality"}},{{"category_id":"f6ff5bc3-aca0-4e4a-8627-e760d0aca78f","category_name":"Experience"}},{{"category_id":"cc771f63-74e7-4c61-882e-3eda9d8ba5d7","category_name":"Price"}}], "classification_instructions": []}}
Assistant:{{"keywords": ["bad service", "slow", "food", "tip", "terrible", "waitresses"],"category_id": "f6ff5bc3-aca0-4e4a-8627-e760d0aca78f","category_name": "Experience"}}
</example>
</example>
### Memory
Here are the chat histories between human and assistant, inside <histories></histories> XML tags.
<histories>

4
api/core/workflow/workflow_entry.py
Unescape Escape View File

@ -9,6 +9,7 @@ from core.app.apps.base_app_queue_manager import GenerateTaskStoppedError
from core.app.entities.app_invoke_entities import InvokeFrom
from core.file.models import File
from core.workflow.callbacks import WorkflowCallback
from core.workflow.constants import ENVIRONMENT_VARIABLE_NODE_ID
from core.workflow.entities.variable_pool import VariablePool
from core.workflow.errors import WorkflowNodeRunFailedError
from core.workflow.graph_engine.entities.event import GraphEngineEvent, GraphRunFailedEvent, InNodeEvent
@ -364,4 +365,5 @@ class WorkflowEntry:
input_value = file_factory.build_from_mappings(mappings=input_value, tenant_id=tenant_id)
# append variable and value to variable pool
variable_pool.add([variable_node_id] + variable_key_list, input_value)
if variable_node_id != ENVIRONMENT_VARIABLE_NODE_ID:
variable_pool.add([variable_node_id] + variable_key_list, input_value)

3
api/docker/entrypoint.sh
Unescape Escape View File

@ -20,7 +20,8 @@ if [[ "${MODE}" == "worker" ]]; then
CONCURRENCY_OPTION="-c ${CELERY_WORKER_AMOUNT:-1}"
fi
exec celery -A app.celery worker -P ${CELERY_WORKER_CLASS:-gevent} $CONCURRENCY_OPTION --loglevel ${LOG_LEVEL:-INFO} \
exec celery -A app.celery worker -P ${CELERY_WORKER_CLASS:-gevent} $CONCURRENCY_OPTION \
--max-tasks-per-child ${MAX_TASK_PRE_CHILD:-50} --loglevel ${LOG_LEVEL:-INFO} \
-Q ${CELERY_QUEUES:-dataset,mail,ops_trace,app_deletion}
elif [[ "${MODE}" == "beat" ]]; then

339
api/extensions/ext_otel.py
Unescape Escape View File

@ -8,192 +8,197 @@ from typing import Union
from celery.signals import worker_init # type: ignore
from flask_login import user_loaded_from_request, user_logged_in # type: ignore
from opentelemetry import trace
from opentelemetry.exporter.otlp.proto.http.metric_exporter import OTLPMetricExporter
from opentelemetry.exporter.otlp.proto.http.trace_exporter import OTLPSpanExporter
from opentelemetry.instrumentation.celery import CeleryInstrumentor
from opentelemetry.instrumentation.flask import FlaskInstrumentor
from opentelemetry.instrumentation.sqlalchemy import SQLAlchemyInstrumentor
from opentelemetry.metrics import get_meter, get_meter_provider, set_meter_provider
from opentelemetry.propagate import set_global_textmap
from opentelemetry.propagators.b3 import B3Format
from opentelemetry.propagators.composite import CompositePropagator
from opentelemetry.sdk.metrics import MeterProvider
from opentelemetry.sdk.metrics.export import ConsoleMetricExporter, PeriodicExportingMetricReader
from opentelemetry.sdk.resources import Resource
from opentelemetry.sdk.trace import TracerProvider
from opentelemetry.sdk.trace.export import (
BatchSpanProcessor,
ConsoleSpanExporter,
)
from opentelemetry.sdk.trace.sampling import ParentBasedTraceIdRatio
from opentelemetry.semconv.resource import ResourceAttributes
from opentelemetry.trace import Span, get_current_span, get_tracer_provider, set_tracer_provider
from opentelemetry.trace.propagation.tracecontext import TraceContextTextMapPropagator
from opentelemetry.trace.status import StatusCode
from configs import dify_config
from dify_app import DifyApp
class ExceptionLoggingHandler(logging.Handler):
"""Custom logging handler that creates spans for logging.exception() calls"""
def emit(self, record):
try:
if record.exc_info:
tracer = get_tracer_provider().get_tracer("dify.exception.logging")
with tracer.start_as_current_span(
"log.exception",
attributes={
"log.level": record.levelname,
"log.message": record.getMessage(),
"log.logger": record.name,
"log.file.path": record.pathname,
"log.file.line": record.lineno,
},
) as span:
span.set_status(StatusCode.ERROR)
span.record_exception(record.exc_info[1])
span.set_attribute("exception.type", record.exc_info[0].__name__)
span.set_attribute("exception.message", str(record.exc_info[1]))
except Exception:
pass
@user_logged_in.connect
@user_loaded_from_request.connect
def on_user_loaded(_sender, user):
if user:
current_span = get_current_span()
if current_span:
current_span.set_attribute("service.tenant.id", user.current_tenant_id)
current_span.set_attribute("service.user.id", user.id)
def init_app(app: DifyApp):
if dify_config.ENABLE_OTEL:
setup_context_propagation()
# Initialize OpenTelemetry
# Follow Semantic Convertions 1.32.0 to define resource attributes
resource = Resource(
attributes={
ResourceAttributes.SERVICE_NAME: dify_config.APPLICATION_NAME,
ResourceAttributes.SERVICE_VERSION: f"dify-{dify_config.CURRENT_VERSION}-{dify_config.COMMIT_SHA}",
ResourceAttributes.PROCESS_PID: os.getpid(),
ResourceAttributes.DEPLOYMENT_ENVIRONMENT: f"{dify_config.DEPLOY_ENV}-{dify_config.EDITION}",
ResourceAttributes.HOST_NAME: socket.gethostname(),
ResourceAttributes.HOST_ARCH: platform.machine(),
"custom.deployment.git_commit": dify_config.COMMIT_SHA,
ResourceAttributes.HOST_ID: platform.node(),
ResourceAttributes.OS_TYPE: platform.system().lower(),
ResourceAttributes.OS_DESCRIPTION: platform.platform(),
ResourceAttributes.OS_VERSION: platform.version(),
}
)
sampler = ParentBasedTraceIdRatio(dify_config.OTEL_SAMPLING_RATE)
provider = TracerProvider(resource=resource, sampler=sampler)
set_tracer_provider(provider)
exporter: Union[OTLPSpanExporter, ConsoleSpanExporter]
metric_exporter: Union[OTLPMetricExporter, ConsoleMetricExporter]
if dify_config.OTEL_EXPORTER_TYPE == "otlp":
exporter = OTLPSpanExporter(
endpoint=dify_config.OTLP_BASE_ENDPOINT + "/v1/traces",
headers={"Authorization": f"Bearer {dify_config.OTLP_API_KEY}"},
)
metric_exporter = OTLPMetricExporter(
endpoint=dify_config.OTLP_BASE_ENDPOINT + "/v1/metrics",
headers={"Authorization": f"Bearer {dify_config.OTLP_API_KEY}"},
)
else:
# Fallback to console exporter
exporter = ConsoleSpanExporter()
metric_exporter = ConsoleMetricExporter()
provider.add_span_processor(
BatchSpanProcessor(
exporter,
max_queue_size=dify_config.OTEL_MAX_QUEUE_SIZE,
schedule_delay_millis=dify_config.OTEL_BATCH_EXPORT_SCHEDULE_DELAY,
max_export_batch_size=dify_config.OTEL_MAX_EXPORT_BATCH_SIZE,
export_timeout_millis=dify_config.OTEL_BATCH_EXPORT_TIMEOUT,
)
)
reader = PeriodicExportingMetricReader(
metric_exporter,
export_interval_millis=dify_config.OTEL_METRIC_EXPORT_INTERVAL,
export_timeout_millis=dify_config.OTEL_METRIC_EXPORT_TIMEOUT,
)
set_meter_provider(MeterProvider(resource=resource, metric_readers=[reader]))
if not is_celery_worker():
init_flask_instrumentor(app)
CeleryInstrumentor(tracer_provider=get_tracer_provider(), meter_provider=get_meter_provider()).instrument()
instrument_exception_logging()
init_sqlalchemy_instrumentor(app)
atexit.register(shutdown_tracer)
from opentelemetry.trace import get_current_span
if user:
current_span = get_current_span()
if current_span:
current_span.set_attribute("service.tenant.id", user.current_tenant_id)
current_span.set_attribute("service.user.id", user.id)
def is_celery_worker():
return "celery" in sys.argv[0].lower()
def init_app(app: DifyApp):
def is_celery_worker():
return "celery" in sys.argv[0].lower()
def instrument_exception_logging():
exception_handler = ExceptionLoggingHandler()
logging.getLogger().addHandler(exception_handler)
def instrument_exception_logging():
exception_handler = ExceptionLoggingHandler()
logging.getLogger().addHandler(exception_handler)
def init_flask_instrumentor(app: DifyApp):
meter = get_meter("http_metrics", version=dify_config.CURRENT_VERSION)
_http_response_counter = meter.create_counter(
"http.server.response.count", description="Total number of HTTP responses by status code", unit="{response}"
)
def response_hook(span: Span, status: str, response_headers: list):
if span and span.is_recording():
if status.startswith("2"):
span.set_status(StatusCode.OK)
else:
span.set_status(StatusCode.ERROR, status)
status = status.split(" ")[0]
status_code = int(status)
status_class = f"{status_code // 100}xx"
_http_response_counter.add(1, {"status_code": status_code, "status_class": status_class})
instrumentor = FlaskInstrumentor()
if dify_config.DEBUG:
logging.info("Initializing Flask instrumentor")
instrumentor.instrument_app(app, response_hook=response_hook)
def init_sqlalchemy_instrumentor(app: DifyApp):
with app.app_context():
engines = list(app.extensions["sqlalchemy"].engines.values())
SQLAlchemyInstrumentor().instrument(enable_commenter=True, engines=engines)
def setup_context_propagation():
# Configure propagators
set_global_textmap(
CompositePropagator(
[
TraceContextTextMapPropagator(), # W3C trace context
B3Format(), # B3 propagation (used by many systems)
]
)
)
def init_flask_instrumentor(app: DifyApp):
meter = get_meter("http_metrics", version=dify_config.CURRENT_VERSION)
_http_response_counter = meter.create_counter(
"http.server.response.count", description="Total number of HTTP responses by status code", unit="{response}"
def shutdown_tracer():
provider = trace.get_tracer_provider()
if hasattr(provider, "force_flush"):
provider.force_flush()
class ExceptionLoggingHandler(logging.Handler):
"""Custom logging handler that creates spans for logging.exception() calls"""
def emit(self, record):
try:
if record.exc_info:
tracer = get_tracer_provider().get_tracer("dify.exception.logging")
with tracer.start_as_current_span(
"log.exception",
attributes={
"log.level": record.levelname,
"log.message": record.getMessage(),
"log.logger": record.name,
"log.file.path": record.pathname,
"log.file.line": record.lineno,
},
) as span:
span.set_status(StatusCode.ERROR)
span.record_exception(record.exc_info[1])
span.set_attribute("exception.type", record.exc_info[0].__name__)
span.set_attribute("exception.message", str(record.exc_info[1]))
except Exception:
pass
from opentelemetry import trace
from opentelemetry.exporter.otlp.proto.http.metric_exporter import OTLPMetricExporter
from opentelemetry.exporter.otlp.proto.http.trace_exporter import OTLPSpanExporter
from opentelemetry.instrumentation.celery import CeleryInstrumentor
from opentelemetry.instrumentation.flask import FlaskInstrumentor
from opentelemetry.instrumentation.sqlalchemy import SQLAlchemyInstrumentor
from opentelemetry.metrics import get_meter, get_meter_provider, set_meter_provider
from opentelemetry.propagate import set_global_textmap
from opentelemetry.propagators.b3 import B3Format
from opentelemetry.propagators.composite import CompositePropagator
from opentelemetry.sdk.metrics import MeterProvider
from opentelemetry.sdk.metrics.export import ConsoleMetricExporter, PeriodicExportingMetricReader
from opentelemetry.sdk.resources import Resource
from opentelemetry.sdk.trace import TracerProvider
from opentelemetry.sdk.trace.export import (
BatchSpanProcessor,
ConsoleSpanExporter,
)
def response_hook(span: Span, status: str, response_headers: list):
if span and span.is_recording():
if status.startswith("2"):
span.set_status(StatusCode.OK)
else:
span.set_status(StatusCode.ERROR, status)
status = status.split(" ")[0]
status_code = int(status)
status_class = f"{status_code // 100}xx"
_http_response_counter.add(1, {"status_code": status_code, "status_class": status_class})
instrumentor = FlaskInstrumentor()
if dify_config.DEBUG:
logging.info("Initializing Flask instrumentor")
instrumentor.instrument_app(app, response_hook=response_hook)
def init_sqlalchemy_instrumentor(app: DifyApp):
with app.app_context():
engines = list(app.extensions["sqlalchemy"].engines.values())
SQLAlchemyInstrumentor().instrument(enable_commenter=True, engines=engines)
def setup_context_propagation():
# Configure propagators
set_global_textmap(
CompositePropagator(
[
TraceContextTextMapPropagator(), # W3C trace context
B3Format(), # B3 propagation (used by many systems)
]
from opentelemetry.sdk.trace.sampling import ParentBasedTraceIdRatio
from opentelemetry.semconv.resource import ResourceAttributes
from opentelemetry.trace import Span, get_tracer_provider, set_tracer_provider
from opentelemetry.trace.propagation.tracecontext import TraceContextTextMapPropagator
from opentelemetry.trace.status import StatusCode
setup_context_propagation()
# Initialize OpenTelemetry
# Follow Semantic Convertions 1.32.0 to define resource attributes
resource = Resource(
attributes={
ResourceAttributes.SERVICE_NAME: dify_config.APPLICATION_NAME,
ResourceAttributes.SERVICE_VERSION: f"dify-{dify_config.CURRENT_VERSION}-{dify_config.COMMIT_SHA}",
ResourceAttributes.PROCESS_PID: os.getpid(),
ResourceAttributes.DEPLOYMENT_ENVIRONMENT: f"{dify_config.DEPLOY_ENV}-{dify_config.EDITION}",
ResourceAttributes.HOST_NAME: socket.gethostname(),
ResourceAttributes.HOST_ARCH: platform.machine(),
"custom.deployment.git_commit": dify_config.COMMIT_SHA,
ResourceAttributes.HOST_ID: platform.node(),
ResourceAttributes.OS_TYPE: platform.system().lower(),
ResourceAttributes.OS_DESCRIPTION: platform.platform(),
ResourceAttributes.OS_VERSION: platform.version(),
}
)
sampler = ParentBasedTraceIdRatio(dify_config.OTEL_SAMPLING_RATE)
provider = TracerProvider(resource=resource, sampler=sampler)
set_tracer_provider(provider)
exporter: Union[OTLPSpanExporter, ConsoleSpanExporter]
metric_exporter: Union[OTLPMetricExporter, ConsoleMetricExporter]
if dify_config.OTEL_EXPORTER_TYPE == "otlp":
exporter = OTLPSpanExporter(
endpoint=dify_config.OTLP_BASE_ENDPOINT + "/v1/traces",
headers={"Authorization": f"Bearer {dify_config.OTLP_API_KEY}"},
)
metric_exporter = OTLPMetricExporter(
endpoint=dify_config.OTLP_BASE_ENDPOINT + "/v1/metrics",
headers={"Authorization": f"Bearer {dify_config.OTLP_API_KEY}"},
)
else:
# Fallback to console exporter
exporter = ConsoleSpanExporter()
metric_exporter = ConsoleMetricExporter()
provider.add_span_processor(
BatchSpanProcessor(
exporter,
max_queue_size=dify_config.OTEL_MAX_QUEUE_SIZE,
schedule_delay_millis=dify_config.OTEL_BATCH_EXPORT_SCHEDULE_DELAY,
max_export_batch_size=dify_config.OTEL_MAX_EXPORT_BATCH_SIZE,
export_timeout_millis=dify_config.OTEL_BATCH_EXPORT_TIMEOUT,
)
)
reader = PeriodicExportingMetricReader(
metric_exporter,
export_interval_millis=dify_config.OTEL_METRIC_EXPORT_INTERVAL,
export_timeout_millis=dify_config.OTEL_METRIC_EXPORT_TIMEOUT,
)
set_meter_provider(MeterProvider(resource=resource, metric_readers=[reader]))
if not is_celery_worker():
init_flask_instrumentor(app)
CeleryInstrumentor(tracer_provider=get_tracer_provider(), meter_provider=get_meter_provider()).instrument()
instrument_exception_logging()
init_sqlalchemy_instrumentor(app)
atexit.register(shutdown_tracer)
@worker_init.connect(weak=False)
def init_celery_worker(*args, **kwargs):
tracer_provider = get_tracer_provider()
metric_provider = get_meter_provider()
if dify_config.DEBUG:
logging.info("Initializing OpenTelemetry for Celery worker")
CeleryInstrumentor(tracer_provider=tracer_provider, meter_provider=metric_provider).instrument()
def is_enabled():
return dify_config.ENABLE_OTEL
def shutdown_tracer():
provider = trace.get_tracer_provider()
if hasattr(provider, "force_flush"):
provider.force_flush()
@worker_init.connect(weak=False)
def init_celery_worker(*args, **kwargs):
if dify_config.ENABLE_OTEL:
from opentelemetry.instrumentation.celery import CeleryInstrumentor
from opentelemetry.metrics import get_meter_provider
from opentelemetry.trace import get_tracer_provider
tracer_provider = get_tracer_provider()
metric_provider = get_meter_provider()
if dify_config.DEBUG:
logging.info("Initializing OpenTelemetry for Celery worker")
CeleryInstrumentor(tracer_provider=tracer_provider, meter_provider=metric_provider).instrument()

63
api/extensions/ext_otel_patch.py
Unescape Escape View File

@ -1,63 +0,0 @@
"""
Patch for OpenTelemetry context detach method to handle None tokens gracefully.
This patch addresses the issue where OpenTelemetry's context.detach() method raises a TypeError
when called with a None token. The error occurs in the contextvars_context.py file where it tries
to call reset() on a None token.
Related GitHub issue: https://github.com/langgenius/dify/issues/18496
Error being fixed:
```
Traceback (most recent call last):
File "opentelemetry/context/__init__.py", line 154, in detach
_RUNTIME_CONTEXT.detach(token)
File "opentelemetry/context/contextvars_context.py", line 50, in detach
self._current_context.reset(token) # type: ignore
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
TypeError: expected an instance of Token, got None
```
Instead of modifying the third-party package directly, this patch monkey-patches the
context.detach method to gracefully handle None tokens.
"""
import logging
from functools import wraps
from opentelemetry import context
logger = logging.getLogger(__name__)
# Store the original detach method
original_detach = context.detach
# Create a patched version that handles None tokens
@wraps(original_detach)
def patched_detach(token):
"""
A patched version of context.detach that handles None tokens gracefully.
"""
if token is None:
logger.debug("Attempted to detach a None token, skipping")
return
return original_detach(token)
def is_enabled():
"""
Check if the extension is enabled.
Always enable this patch to prevent errors even when OpenTelemetry is disabled.
"""
return True
def init_app(app):
"""
Initialize the OpenTelemetry context patch.
"""
# Replace the original detach method with our patched version
context.detach = patched_detach
logger.info("OpenTelemetry context.detach patched to handle None tokens")

4
api/migrations/versions/2024_10_10_0516-bbadea11becb_add_name_and_size_to_tool_files.py
Unescape Escape View File

@ -21,14 +21,14 @@ def upgrade():
# ### commands auto generated by Alembic - please adjust! ###
# Get the database connection
conn = op.get_bind()
# Use SQLAlchemy inspector to get the columns of the 'tool_files' table
inspector = sa.inspect(conn)
columns = [col['name'] for col in inspector.get_columns('tool_files')]
# If 'name' or 'size' columns already exist, exit the upgrade function
if 'name' in columns or 'size' in columns:
return
return
with op.batch_alter_table('tool_files', schema=None) as batch_op:
batch_op.add_column(sa.Column('name', sa.String(), nullable=True))

2
api/migrations/versions/2024_12_20_0628-e1944c35e15e_add_retry_index_field_to_node_execution_.py
Unescape Escape View File

@ -35,4 +35,4 @@ def downgrade():
# batch_op.drop_column('retry_index')
pass
# ### end Alembic commands ###
# ### end Alembic commands ###

2
api/migrations/versions/2024_12_23_1154-d7999dfa4aae_remove_workflow_node_executions_retry_.py
Unescape Escape View File

@ -23,7 +23,7 @@ def upgrade():
conn = op.get_bind()
inspector = inspect(conn)
has_column = 'retry_index' in [col['name'] for col in inspector.get_columns('workflow_node_executions')]
if has_column:
with op.batch_alter_table('workflow_node_executions', schema=None) as batch_op:
batch_op.drop_column('retry_index')

2
api/migrations/versions/64b051264f32_init.py
Unescape Escape View File

@ -1,7 +1,7 @@
"""init
Revision ID: 64b051264f32
Revises:
Revises:
Create Date: 2023-05-13 14:26:59.085018
"""

4
api/migrations/versions/de95f5c77138_migration_serpapi_api_key.py
Unescape Escape View File

@ -99,12 +99,12 @@ def upgrade():
id=id,
tenant_id=tenant_id,
user_id=user_id,
provider='google',
provider='google',
encrypted_credentials=encrypted_credentials,
created_at=created_at,
updated_at=updated_at
)
# ### end Alembic commands ###

10
api/pyproject.toml
Unescape Escape View File

@ -1,6 +1,6 @@
[project]
name = "dify-api"
version = "1.3.0"
dynamic = ["version"]
requires-python = ">=3.11,<3.13"
dependencies = [
@ -81,15 +81,19 @@ dependencies = [
"tokenizers~=0.15.0",
"transformers~=4.35.0",
"unstructured[docx,epub,md,ppt,pptx]~=0.16.1",
"validators==0.21.0",
"weave~=0.51.34",
"yarl~=1.18.3",
"webvtt-py~=0.5.1",
]
# Before adding new dependency, consider place it in
# alphabet order (a-z) and suitable group.
[tool.setuptools]
packages = []
[tool.uv]
default-groups = ["storage", "tools", "vdb"]
package = false
[dependency-groups]
@ -191,6 +195,6 @@ vdb = [
"tidb-vector==0.0.9",
"upstash-vector==0.6.0",
"volcengine-compat~=1.0.156",
"weaviate-client~=3.21.0",
"weaviate-client~=3.24.0",
"xinference-client~=1.2.2",
]

6
api/services/external_knowledge_service.py
Unescape Escape View File

@ -2,9 +2,9 @@ import json
from copy import deepcopy
from datetime import UTC, datetime
from typing import Any, Optional, Union, cast
from urllib.parse import urlparse
import httpx
import validators
from constants import HIDDEN_VALUE
from core.helper import ssrf_proxy
@ -72,7 +72,9 @@ class ExternalDatasetService:
endpoint = f"{settings['endpoint']}/retrieval"
api_key = settings["api_key"]
if not validators.url(endpoint, simple_host=True):
parsed_url = urlparse(endpoint)
if not all([parsed_url.scheme, parsed_url.netloc]):
if not endpoint.startswith("http://") and not endpoint.startswith("https://"):
raise ValueError(f"invalid endpoint: {endpoint} must start with http:// or https://")
else:

16
api/services/plugin/data_migration.py
Unescape Escape View File

@ -86,9 +86,9 @@ limit 1000"""
update_retrieval_model_sql = ", retrieval_model = :retrieval_model"
params["retrieval_model"] = json.dumps(retrieval_model)
sql = f"""update {table_name}
set {provider_column_name} =
concat('{DEFAULT_PLUGIN_ID}/', {provider_column_name}, '/', {provider_column_name})
sql = f"""update {table_name}
set {provider_column_name} =
concat('{DEFAULT_PLUGIN_ID}/', {provider_column_name}, '/', {provider_column_name})
{update_retrieval_model_sql}
where id = :record_id"""
conn.execute(db.text(sql), params)
@ -131,10 +131,10 @@ limit 1000"""
while True:
sql = f"""
SELECT id, {provider_column_name} AS provider_name
SELECT id, {provider_column_name} AS provider_name
FROM {table_name}
WHERE {provider_column_name} NOT LIKE '%/%'
AND {provider_column_name} IS NOT NULL
WHERE {provider_column_name} NOT LIKE '%/%'
AND {provider_column_name} IS NOT NULL
AND {provider_column_name} != ''
AND id > :last_id
ORDER BY id ASC
@ -183,8 +183,8 @@ limit 1000"""
if batch_updates:
update_sql = f"""
UPDATE {table_name}
SET {provider_column_name} = :updated_value
UPDATE {table_name}
SET {provider_column_name} = :updated_value
WHERE id = :record_id
"""
conn.execute(db.text(update_sql), [{"updated_value": u, "record_id": r} for u, r in batch_updates])

4
api/templates/clean_document_job_mail_template-US.html
Unescape Escape View File

@ -77,7 +77,7 @@
<h1>Some Documents in Your Knowledge Base Have Been Disabled</h1>
<p>Dear {{userName}},</p>
<p>
We're sorry for the inconvenience. To ensure optimal performance, documents
We're sorry for the inconvenience. To ensure optimal performance, documents
that haven’t been updated or accessed in the past 30 days have been disabled in
your knowledge bases:
</p>
@ -97,4 +97,4 @@
</div>
</div>
</body>
</html>
</html>

2
api/templates/delete_account_code_email_template_en-US.html
Unescape Escape View File

@ -122,4 +122,4 @@
</div>
</body>
</html>
</html>

2
api/templates/delete_account_success_template_en-US.html
Unescape Escape View File

@ -102,4 +102,4 @@
</div>
</body>
</html>
</html>

2
api/tests/integration_tests/.gitignore vendored
Unescape Escape View File

@ -1 +1 @@
.env.test
.env.test

59
api/tests/integration_tests/vdb/opensearch/test_opensearch.py
Unescape Escape View File

@ -23,13 +23,70 @@ def setup_mock_redis():
ext_redis.redis_client.lock = MagicMock(return_value=mock_redis_lock)
class TestOpenSearchConfig:
def test_to_opensearch_params(self):
config = OpenSearchConfig(
host="localhost",
port=9200,
secure=True,
user="admin",
password="password",
)
params = config.to_opensearch_params()
assert params["hosts"] == [{"host": "localhost", "port": 9200}]
assert params["use_ssl"] is True
assert params["verify_certs"] is True
assert params["connection_class"].__name__ == "Urllib3HttpConnection"
assert params["http_auth"] == ("admin", "password")
@patch("boto3.Session")
@patch("core.rag.datasource.vdb.opensearch.opensearch_vector.Urllib3AWSV4SignerAuth")
def test_to_opensearch_params_with_aws_managed_iam(
self, mock_aws_signer_auth: MagicMock, mock_boto_session: MagicMock
):
mock_credentials = MagicMock()
mock_boto_session.return_value.get_credentials.return_value = mock_credentials
mock_auth_instance = MagicMock()
mock_aws_signer_auth.return_value = mock_auth_instance
aws_region = "ap-southeast-2"
aws_service = "aoss"
host = f"aoss-endpoint.{aws_region}.aoss.amazonaws.com"
port = 9201
config = OpenSearchConfig(
host=host,
port=port,
secure=True,
auth_method="aws_managed_iam",
aws_region=aws_region,
aws_service=aws_service,
)
params = config.to_opensearch_params()
assert params["hosts"] == [{"host": host, "port": port}]
assert params["use_ssl"] is True
assert params["verify_certs"] is True
assert params["connection_class"].__name__ == "Urllib3HttpConnection"
assert params["http_auth"] is mock_auth_instance
mock_aws_signer_auth.assert_called_once_with(
credentials=mock_credentials, region=aws_region, service=aws_service
)
assert mock_boto_session.return_value.get_credentials.called
class TestOpenSearchVector:
def setup_method(self):
self.collection_name = "test_collection"
self.example_doc_id = "example_doc_id"
self.vector = OpenSearchVector(
collection_name=self.collection_name,
config=OpenSearchConfig(host="localhost", port=9200, user="admin", password="password", secure=False),
config=OpenSearchConfig(host="localhost", port=9200, secure=False, user="admin", password="password"),
)
self.vector._client = MagicMock()

2
api/tests/unit_tests/.gitignore vendored
Unescape Escape View File

@ -1 +1 @@
.env.test
.env.test

7
api/tests/unit_tests/core/workflow/graph_engine/test_graph_engine.py
Unescape Escape View File

@ -864,10 +864,11 @@ def test_condition_parallel_correct_output(mock_close, mock_remove, app):
with patch.object(CodeNode, "_run", new=code_generator):
generator = graph_engine.run()
stream_content = ""
res_content = "VAT:\ndify 123"
wrong_content = ["Stamp Duty", "other"]
for item in generator:
if isinstance(item, NodeRunStreamChunkEvent):
stream_content += f"{item.chunk_content}\n"
if isinstance(item, GraphRunSucceededEvent):
assert item.outputs == {"answer": res_content}
assert stream_content == res_content + "\n"
assert item.outputs is not None
answer = item.outputs["answer"]
assert all(rc not in answer for rc in wrong_content)

29
api/uv.lock
Unescape Escape View File

@ -1155,7 +1155,6 @@ wheels = [
[[package]]
name = "dify-api"
version = "1.3.0"
source = { virtual = "." }
dependencies = [
{ name = "authlib" },
@ -1233,8 +1232,8 @@ dependencies = [
{ name = "tokenizers" },
{ name = "transformers" },
{ name = "unstructured", extra = ["docx", "epub", "md", "ppt", "pptx"] },
{ name = "validators" },
{ name = "weave" },
{ name = "webvtt-py" },
{ name = "yarl" },
]
@ -1403,8 +1402,8 @@ requires-dist = [
{ name = "tokenizers", specifier = "~=0.15.0" },
{ name = "transformers", specifier = "~=4.35.0" },
{ name = "unstructured", extras = ["docx", "epub", "md", "ppt", "pptx"], specifier = "~=0.16.1" },
{ name = "validators", specifier = "==0.21.0" },
{ name = "weave", specifier = "~=0.51.34" },
{ name = "webvtt-py", specifier = "~=0.5.1" },
{ name = "yarl", specifier = "~=1.18.3" },
]
@ -1492,7 +1491,7 @@ vdb = [
{ name = "tidb-vector", specifier = "==0.0.9" },
{ name = "upstash-vector", specifier = "==0.6.0" },
{ name = "volcengine-compat", specifier = "~=1.0.156" },
{ name = "weaviate-client", specifier = "~=3.21.0" },
{ name = "weaviate-client", specifier = "~=3.24.0" },
{ name = "xinference-client", specifier = "~=1.2.2" },
]
@ -6086,11 +6085,11 @@ wheels = [
[[package]]
name = "validators"
version = "0.21.0"
version = "0.34.0"
source = { registry = "https://pypi.org/simple" }
sdist = { url = "https://files.pythonhosted.org/packages/1f/c5/4095e7a5a6fecc2eca953ad058a3609135d833f986f84951f7e26790d651/validators-0.21.0.tar.gz", hash = "sha256:245b98ab778ed9352a7269c6a8f6c2a839bed5b2a7e3e60273ce399d247dd4b3", size = 20937 }
sdist = { url = "https://files.pythonhosted.org/packages/64/07/91582d69320f6f6daaf2d8072608a4ad8884683d4840e7e4f3a9dbdcc639/validators-0.34.0.tar.gz", hash = "sha256:647fe407b45af9a74d245b943b18e6a816acf4926974278f6dd617778e1e781f", size = 70955 }
wheels = [
{ url = "https://files.pythonhosted.org/packages/ad/50/18dbf2ac594234ee6249bfe3425fa424c18eeb96f29dcd47f199ed6c51bc/validators-0.21.0-py3-none-any.whl", hash = "sha256:3470db6f2384c49727ee319afa2e97aec3f8fad736faa6067e0fd7f9eaf2c551", size = 27686 },
{ url = "https://files.pythonhosted.org/packages/6e/78/36828a4d857b25896f9774c875714ba4e9b3bc8a92d2debe3f4df3a83d4f/validators-0.34.0-py3-none-any.whl", hash = "sha256:c804b476e3e6d3786fa07a30073a4ef694e617805eb1946ceee3fe5a9b8b1321", size = 43536 },
]
[[package]]
@ -6220,17 +6219,16 @@ wheels = [
[[package]]
name = "weaviate-client"
version = "3.21.0"
version = "3.24.2"
source = { registry = "https://pypi.org/simple" }
dependencies = [
{ name = "authlib" },
{ name = "requests" },
{ name = "tqdm" },
{ name = "validators" },
]
sdist = { url = "https://files.pythonhosted.org/packages/b4/a5/c6777a8507249d7a63f4f5d9696eb5f45beac87db0eddfa4438d408cc3b4/weaviate-client-3.21.0.tar.gz", hash = "sha256:ec94ac554883c765e94da8b2947c4f0fa4a0378ed3bbe9f3653df3a5b1745a6d", size = 186970 }
sdist = { url = "https://files.pythonhosted.org/packages/1f/c1/3285a21d8885f2b09aabb65edb9a8e062a35c2d7175e1bb024fa096582ab/weaviate-client-3.24.2.tar.gz", hash = "sha256:6914c48c9a7e5ad0be9399271f9cb85d6f59ab77476c6d4e56a3925bf149edaa", size = 199332 }
wheels = [
{ url = "https://files.pythonhosted.org/packages/df/5b/57b55ad36eb071b57e79f1ea7fba5bfe6a2fe49702607f56726569665d60/weaviate_client-3.21.0-py3-none-any.whl", hash = "sha256:420444ded7106fb000f4f8b2321b5f5fa2387825aa7a303d702accf61026f9d2", size = 99944 },
{ url = "https://files.pythonhosted.org/packages/ab/98/3136d05f93e30cf29e1db280eaadf766df18d812dfe7994bcced653b2340/weaviate_client-3.24.2-py3-none-any.whl", hash = "sha256:bc50ca5fcebcd48de0d00f66700b0cf7c31a97c4cd3d29b4036d77c5d1d9479b", size = 107968 },
]
[[package]]
@ -6282,6 +6280,15 @@ wheels = [
{ url = "https://files.pythonhosted.org/packages/7b/c8/d529f8a32ce40d98309f4470780631e971a5a842b60aec864833b3615786/websockets-14.2-py3-none-any.whl", hash = "sha256:7a6ceec4ea84469f15cf15807a747e9efe57e369c384fa86e022b3bea679b79b", size = 157416 },
]
[[package]]
name = "webvtt-py"
version = "0.5.1"
source = { registry = "https://pypi.org/simple" }
sdist = { url = "https://files.pythonhosted.org/packages/5e/f6/7c9c964681fb148e0293e6860108d378e09ccab2218f9063fd3eb87f840a/webvtt-py-0.5.1.tar.gz", hash = "sha256:2040dd325277ddadc1e0c6cc66cbc4a1d9b6b49b24c57a0c3364374c3e8a3dc1", size = 55128 }
wheels = [
{ url = "https://files.pythonhosted.org/packages/f3/ed/aad7e0f5a462d679f7b4d2e0d8502c3096740c883b5bbed5103146480937/webvtt_py-0.5.1-py3-none-any.whl", hash = "sha256:9d517d286cfe7fc7825e9d4e2079647ce32f5678eb58e39ef544ffbb932610b7", size = 19802 },
]
[[package]]
name = "werkzeug"
version = "3.1.3"

2
dev/pytest/pytest_all_tests.sh
Unescape Escape View File

@ -11,4 +11,4 @@ dev/pytest/pytest_tools.sh
dev/pytest/pytest_workflow.sh
# Unit tests
dev/pytest/pytest_unit_tests.sh
dev/pytest/pytest_unit_tests.sh

2
dev/pytest/pytest_model_runtime.sh
Unescape Escape View File

@ -10,4 +10,4 @@ pytest api/tests/integration_tests/model_runtime/anthropic \
api/tests/integration_tests/model_runtime/fireworks \
api/tests/integration_tests/model_runtime/nomic \
api/tests/integration_tests/model_runtime/mixedbread \
api/tests/integration_tests/model_runtime/voyage
api/tests/integration_tests/model_runtime/voyage

12
docker/.env.example
Unescape Escape View File

@ -39,6 +39,12 @@ APP_WEB_URL=
# File preview or download Url prefix.
# used to display File preview or download Url to the front-end or as Multi-model inputs;
# Url is signed and has expiration time.
# Setting FILES_URL is required for file processing plugins.
# - For https://example.com, use FILES_URL=https://example.com
# - For http://example.com, use FILES_URL=http://example.com
# Recommendation: use a dedicated domain (e.g., https://upload.example.com).
# Alternatively, use http://<your-ip>:5001 or http://api:5001,
# ensuring port 5001 is externally accessible (see docker-compose.yaml).
FILES_URL=
# ------------------------------
@ -520,9 +526,13 @@ RELYT_DATABASE=postgres
# open search configuration, only available when VECTOR_STORE is `opensearch`
OPENSEARCH_HOST=opensearch
OPENSEARCH_PORT=9200
OPENSEARCH_SECURE=true
OPENSEARCH_AUTH_METHOD=basic
OPENSEARCH_USER=admin
OPENSEARCH_PASSWORD=admin
OPENSEARCH_SECURE=true
# If using AWS managed IAM, e.g. Managed Cluster or OpenSearch Serverless
OPENSEARCH_AWS_REGION=ap-southeast-1
OPENSEARCH_AWS_SERVICE=aoss
# tencent vector configurations, only available when VECTOR_STORE is `tencent`
TENCENT_VECTOR_DB_URL=http://127.0.0.1

1
docker/README.md
Unescape Escape View File

@ -14,7 +14,6 @@ Welcome to the new `docker` directory for deploying Dify using Docker Compose. T
- **Unified Vector Database Services**: All vector database services are now managed from a single Docker Compose file `docker-compose.yaml`. You can switch between different vector databases by setting the `VECTOR_STORE` environment variable in your `.env` file.
- **Mandatory .env File**: A `.env` file is now required to run `docker compose up`. This file is crucial for configuring your deployment and for any custom settings to persist through upgrades.
- **Legacy Support**: Previous deployment files are now located in the `docker-legacy` directory and will no longer be maintained.
### How to Deploy Dify with `docker-compose.yaml`

4
docker/couchbase-server/Dockerfile
Unescape Escape View File

@ -1,4 +1,4 @@
FROM couchbase/server:latest AS stage_base
# FROM couchbase:latest AS stage_base
# FROM couchbase:latest AS stage_base
COPY init-cbserver.sh /opt/couchbase/init/
RUN chmod +x /opt/couchbase/init/init-cbserver.sh
RUN chmod +x /opt/couchbase/init/init-cbserver.sh

8
docker/couchbase-server/init-cbserver.sh
Unescape Escape View File

@ -1,8 +1,8 @@
#!/bin/bash
# used to start couchbase server - can't get around this as docker compose only allows you to start one command - so we have to start couchbase like the standard couchbase Dockerfile would
# used to start couchbase server - can't get around this as docker compose only allows you to start one command - so we have to start couchbase like the standard couchbase Dockerfile would
# https://github.com/couchbase/docker/blob/master/enterprise/couchbase-server/7.2.0/Dockerfile#L88
/entrypoint.sh couchbase-server &
/entrypoint.sh couchbase-server &
# track if setup is complete so we don't try to setup again
FILE=/opt/couchbase/init/setupComplete.txt
@ -36,9 +36,9 @@ if ! [ -f "$FILE" ]; then
--bucket-ramsize $COUCHBASE_BUCKET_RAMSIZE \
--bucket-type couchbase
# create file so we know that the cluster is setup and don't run the setup again
# create file so we know that the cluster is setup and don't run the setup again
touch $FILE
fi
fi
# docker compose will stop the container from running unless we do this
# known issue and workaround
tail -f /dev/null

6
docker/docker-compose-template.yaml
Unescape Escape View File

@ -2,7 +2,7 @@ x-shared-env: &shared-api-worker-env
services:
# API service
api:
image: langgenius/dify-api:1.3.0
image: langgenius/dify-api:1.3.1
restart: always
environment:
# Use the shared environment variables.
@ -31,7 +31,7 @@ services:
# worker service
# The Celery worker for processing the queue.
worker:
image: langgenius/dify-api:1.3.0
image: langgenius/dify-api:1.3.1
restart: always
environment:
# Use the shared environment variables.
@ -57,7 +57,7 @@ services:
# Frontend web application.
web:
image: langgenius/dify-web:1.3.0
image: langgenius/dify-web:1.3.1
restart: always
environment:
CONSOLE_API_URL: ${CONSOLE_API_URL:-}

BIN
docker/docker-compose.png
View File

Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 62 KiB

After

Width:  |  Height:  |  Size: 170 KiB

11
docker/docker-compose.yaml
Unescape Escape View File

@ -225,9 +225,12 @@ x-shared-env: &shared-api-worker-env
RELYT_DATABASE: ${RELYT_DATABASE:-postgres}
OPENSEARCH_HOST: ${OPENSEARCH_HOST:-opensearch}
OPENSEARCH_PORT: ${OPENSEARCH_PORT:-9200}
OPENSEARCH_SECURE: ${OPENSEARCH_SECURE:-true}
OPENSEARCH_AUTH_METHOD: ${OPENSEARCH_AUTH_METHOD:-basic}
OPENSEARCH_USER: ${OPENSEARCH_USER:-admin}
OPENSEARCH_PASSWORD: ${OPENSEARCH_PASSWORD:-admin}
OPENSEARCH_SECURE: ${OPENSEARCH_SECURE:-true}
OPENSEARCH_AWS_REGION: ${OPENSEARCH_AWS_REGION:-ap-southeast-1}
OPENSEARCH_AWS_SERVICE: ${OPENSEARCH_AWS_SERVICE:-aoss}
TENCENT_VECTOR_DB_URL: ${TENCENT_VECTOR_DB_URL:-http://127.0.0.1}
TENCENT_VECTOR_DB_API_KEY: ${TENCENT_VECTOR_DB_API_KEY:-dify}
TENCENT_VECTOR_DB_TIMEOUT: ${TENCENT_VECTOR_DB_TIMEOUT:-30}
@ -488,7 +491,7 @@ x-shared-env: &shared-api-worker-env
services:
# API service
api:
image: langgenius/dify-api:1.3.0
image: langgenius/dify-api:1.3.1
restart: always
environment:
# Use the shared environment variables.
@ -517,7 +520,7 @@ services:
# worker service
# The Celery worker for processing the queue.
worker:
image: langgenius/dify-api:1.3.0
image: langgenius/dify-api:1.3.1
restart: always
environment:
# Use the shared environment variables.
@ -543,7 +546,7 @@ services:
# Frontend web application.
web:
image: langgenius/dify-web:1.3.0
image: langgenius/dify-web:1.3.1
restart: always
environment:
CONSOLE_API_URL: ${CONSOLE_API_URL:-}

2
docker/middleware.env.example
Unescape Escape View File

@ -144,4 +144,4 @@ PLUGIN_AZURE_BLOB_STORAGE_CONNECTION_STRING=
# Plugin oss tencent cos
PLUGIN_TENCENT_COS_SECRET_KEY=
PLUGIN_TENCENT_COS_SECRET_ID=
PLUGIN_TENCENT_COS_REGION=
PLUGIN_TENCENT_COS_REGION=

2
docker/nginx/docker-entrypoint.sh
Unescape Escape View File

@ -39,4 +39,4 @@ envsubst "$env_vars" < /etc/nginx/proxy.conf.template > /etc/nginx/proxy.conf
envsubst "$env_vars" < /etc/nginx/conf.d/default.conf.template > /etc/nginx/conf.d/default.conf
# Start Nginx using the default entrypoint
exec nginx -g 'daemon off;'
exec nginx -g 'daemon off;'

2
docker/nginx/https.conf.template
Unescape Escape View File

@ -6,4 +6,4 @@ ssl_certificate_key ${SSL_CERTIFICATE_KEY_PATH};
ssl_protocols ${NGINX_SSL_PROTOCOLS};
ssl_prefer_server_ciphers on;
ssl_session_cache shared:SSL:10m;
ssl_session_timeout 10m;
ssl_session_timeout 10m;

2
docker/nginx/nginx.conf.template
Unescape Escape View File

@ -31,4 +31,4 @@ http {
client_max_body_size ${NGINX_CLIENT_MAX_BODY_SIZE};
include /etc/nginx/conf.d/*.conf;
}
}

4
docker/ssrf_proxy/squid.conf.template
Unescape Escape View File

@ -44,7 +44,7 @@ refresh_pattern . 0 20% 4320
# cache_dir ufs /var/spool/squid 100 16 256
# upstream proxy, set to your own upstream proxy IP to avoid SSRF attacks
# cache_peer 172.1.1.1 parent 3128 0 no-query no-digest no-netdb-exchange default
# cache_peer 172.1.1.1 parent 3128 0 no-query no-digest no-netdb-exchange default
################################## Reverse Proxy To Sandbox ################################
http_port ${REVERSE_PROXY_PORT} accel vhost
@ -53,4 +53,4 @@ acl src_all src all
http_access allow src_all
# Unless the option's size is increased, an error will occur when uploading more than two files.
client_request_buffer_max_size 100 MB
client_request_buffer_max_size 100 MB

2
docker/startupscripts/init.sh
Unescape Escape View File

@ -8,6 +8,6 @@ if [ -f ${DB_INITIALIZED} ]; then
exit
else
echo 'File does not exist. Standards for first time Start up this DB'
"$ORACLE_HOME"/bin/sqlplus -s "/ as sysdba" @"/opt/oracle/scripts/startup/init_user.script";
"$ORACLE_HOME"/bin/sqlplus -s "/ as sysdba" @"/opt/oracle/scripts/startup/init_user.script";
touch ${DB_INITIALIZED}
fi

2
docker/startupscripts/init_user.script
Unescape Escape View File

@ -1,5 +1,5 @@
show pdbs;
ALTER SYSTEM SET PROCESSES=500 SCOPE=SPFILE;
ALTER SYSTEM SET PROCESSES=500 SCOPE=SPFILE;
alter session set container= freepdb1;
create user dify identified by dify DEFAULT TABLESPACE users quota unlimited on users;
grant DB_DEVELOPER_ROLE to dify;

2
docker/tidb/config/pd.toml
Unescape Escape View File

@ -1,4 +1,4 @@
# PD Configuration File reference:
# https://docs.pingcap.com/tidb/stable/pd-configuration-file#pd-configuration-file
[replication]
max-replicas = 1
max-replicas = 1

2
docker/volumes/myscale/config/users.d/custom_users_config.xml
Unescape Escape View File

@ -14,4 +14,4 @@
<access_management>1</access_management>
</default>
</users>
</clickhouse>
</clickhouse>

2
docker/volumes/oceanbase/init.d/vec_memory.sql
Unescape Escape View File

@ -1 +1 @@
ALTER SYSTEM SET ob_vector_memory_limit_percentage = 30;
ALTER SYSTEM SET ob_vector_memory_limit_percentage = 30;

2
sdks/nodejs-client/.gitignore vendored
Unescape Escape View File

@ -45,4 +45,4 @@ package-lock.json
.yarnrc.yml
# pmpm
pnpm-lock.yaml
pnpm-lock.yaml

12
sdks/nodejs-client/index.d.ts vendored
Unescape Escape View File

@ -26,7 +26,7 @@ export declare class DifyClient {
params?: Params,
stream?: boolean,
headerParams?: HeaderParams
): Promise<any>;
): Promise<any>;
messageFeedback(message_id: string, rating: number, user: User): Promise<any>;
@ -64,9 +64,9 @@ export declare class ChatClient extends DifyClient {
getConversations(
user: User,
first_id?: string | null,
limit?: number | null,
user: User,
first_id?: string | null,
limit?: number | null,
pinned?: boolean | null
): Promise<any>;
@ -80,7 +80,7 @@ export declare class ChatClient extends DifyClient {
renameConversation(conversation_id: string, name: string, user: User,auto_generate:boolean): Promise<any>;
deleteConversation(conversation_id: string, user: User): Promise<any>;
audioToText(data: FormData): Promise<any>;
}
@ -88,4 +88,4 @@ export declare class WorkflowClient extends DifyClient {
run(inputs: any, user: User, stream?: boolean,): Promise<any>;
stop(task_id: string, user: User): Promise<any>;
}
}

10
sdks/nodejs-client/index.js
Unescape Escape View File

@ -334,12 +334,12 @@ export class ChatClient extends DifyClient {
export class WorkflowClient extends DifyClient {
run(inputs,user,stream) {
const data = {
inputs,
const data = {
inputs,
response_mode: stream ? "streaming" : "blocking",
user
user
};
return this.sendRequest(
routes.runWorkflow.method,
routes.runWorkflow.url(),
@ -357,4 +357,4 @@ export class WorkflowClient extends DifyClient {
data
);
}
}
}

2
sdks/nodejs-client/index.test.js
Unescape Escape View File

@ -62,4 +62,4 @@ describe('Send Requests', () => {
errorMessage
)
})
})
})

Some files were not shown because too many files have changed in this diff Show More

Write Preview
Loading…
Cancel
Save