1.0.0 Discussion

[crazywoola]

Self Checks

• I have searched for existing issues search for existing issues, including closed ones.
• I confirm that I am using English to submit this report (我已阅读并同意 Language Policy).
• [FOR CHINESE USERS] 请务必使用英文提交 Issue，否则会被关闭。谢谢！:）
• Please do not modify this template :) and fill in all the required fields.

Content

Before we launch the version 1.0.0, you can talk about the features and bugs here, please attach an issue/pull request link here.

This milestone gathers some bugs and pull requests which related to version 1.0.0.

[tipani86]

Is the team going to share an approximate timeline for SaaS to be upgraded, so us devs can mentally prepare for something to go wrong and be ready for issuing hotfixes?

We have multiple systems running in production hooked on Dify SaaS and in the past even a minor version upgrade might have thrown something off, so we hope the best but anticipate the worst with a major one like 1.0.

[DylanWu92]

Traceback (most recent call last): File "/var/sandbox/sandbox-python/tmp/c00f1c7e_d033_40e8_a1ad_f4f86eea5131.py",
line 48, in File "", line 3, in ModuleNotFoundError: No module named 'pymysql' error: exit status 255
这个怎么解决

[liuhVIP]

AIP access, does not support associated knowledge base?

[Ynewtime]

Kindly review this issue #12715 again, thanks:

[fenghaoyu16]

What's the problem

[onedot]

Aliyu's computing nest is still 0.15.3. How to update to 1.0.0 in the computing nest or when will the computing nest be upgraded to 1.0.0?

[jackerly717]

models in local nor qwen api ,neither can be added to 1.0.0,error shows internel server error

[Gerry-chen2024]

same.....

[Gerry-chen2024]

But, it has been ok, I didn't do anything, and it becomes ok suddenly.

[huangyang-daye]

me too\\\

[huangyang-daye]

yesterday I locally updated dify to 1.0.1, and it succeeded suddenly

[tgh4821]

what's the peoblem? I download the total file on github by zip before and run them run docker. Now, I want to update to 1.0.0.

[NullUserSIe]

You can never get the full repo when you just download the zip from github.

You should git clone the repo from github, or via any github mirrors (gitclone.com?). Only the full repo with .git folder would grant you this ability to checkout to another branch. BTW if you think clone would take you too much time, try using --depth=1.

[tgh4821]

You can never get the full repo when you just download the zip from github.当您仅从 github 下载 zip 时，您永远无法获得完整的 repo。

You should git clone the repo from github, or via any github mirrors (gitclone.com?). Only the full repo with .git folder would grant you this ability to checkout to another branch. BTW if you think clone would take you too much time, try using --depth=1.

It's okay. If there is a tutorial available that can help me understand how to use 'git clone' to obtain the .git folder, that would be great. After all, thanks.

[NullUserSIe]

omitted
Just follow the instructions in the Quick Start and follow the steps:

• Make sure you have already installed Git, be it PortableGit or normal git - you should be able to check it by input git --version and enter in your command prompts (for windows i suggest you to use cmd instead of powershell).
  • if you have already installed it, it should show up with something like:

  D:\Syphurith>git --version
  git version 2.47.1.windows.2
  

  • if not, make sure you've installed git via winget or dkpg or simply download from git website, and have it in your PATH.
• Once comfirmed its existence, select a link of dify from its github repo. And pick the link by simply click on the green button Code from the page. for example..
  • clone from official github: git clone https://github.com/langgenius/dify.git
  • or for chinese they should select another if github is slow: git clone https://gitclone.com/github.com/langgenius/dify.git
• Once you have selected the link for your local repo, do open your command prompt, and direct to the place you want your dify folder. Then, just do the clone, for example: Paste git clone https://gitclone.com/github.com/langgenius/dify.git into your command prompt, and enter.
• Cloning the repo should take you some time, but not too long. There would be a dify folder with a hidden .git folder inside.
• Once it finished and you get a complete repo, use git checkout command if you need.

[tgh4821]

omitted
Just follow the instructions in the Quick Start and follow the steps:

• Make sure you have already installed Git, be it PortableGit or normal git - you should be able to check it by input git --version and enter in your command prompts (for windows i suggest you to use cmd instead of powershell).

  • if you have already installed it, it should show up with something like:

  D:\Syphurith>git --version
  git version 2.47.1.windows.2
  

  • if not, make sure you've installed git via winget or dkpg or simply download from git website, and have it in your PATH.

• Once comfirmed its existence, select a link of dify from its github repo. And pick the link by simply click on the green button Code from the page. for example..

  • clone from official github: git clone https://github.com/langgenius/dify.git
  • or for chinese they should select another if github is slow: git clone https://gitclone.com/github.com/langgenius/dify.git

• Once you have selected the link for your local repo, do open your command prompt, and direct to the place you want your dify folder. Then, just do the clone, for example: Paste git clone https://gitclone.com/github.com/langgenius/dify.git into your command prompt, and enter.

• Cloning the repo should take you some time, but not too long. There would be a dify folder with a hidden .git folder inside.

• Once it finished and you get a complete repo, use git checkout command if you need.

Wow, a so quick answer. I had done it and solved the problem. Thank you again. I hope you have a good day.

[zhangboxushj]

我想请问一下更新到1.0.0之后这个问题怎么解决，各位大佬

[NullUserSIe]

如果没看错，意思大概是：下载ComfyUI插件时达到最大重试次数(3)。你需要检查你的网络。

---

And please be aware you should better pop up your questions in English following the rules next time, thus there may be more guys can help you out.

[NullUserSIe]

Using Dify Cloud (Version 1.0.0), I faced some problems:

---

Problem about management of Models from Online Model Providers

• I have applied for an OpenRouter account, and paid for some credits to make its API working to pass the error 402.
• When i applied my API key to Dify and finally made it work, i find lots of paid models sumed to 36 of those appears under the list.
• I add some models i need and i can not get the list to above 38. and i can not remove any models that automatically filled in the list.

• When dify validates the API from OpenRouter, it even cause a little fee for accessing GPT 3.5 Turbo, and this is without any warnings. Luckily i have some credits or else this might even get my account down.
  

Conclusion:

• I can not remove any models Dify automatically added to the list of OpenRouter.
• I can only add up to 2 models i specified. The last one added will overwrite the first one.
• I don't know the ones automatically added without my confirmation, will even charge me (even that is $0.00001)

These must be bugs, as you can easily manage them if you are using Ollama.

Anyone knows how to remove a certain model from the model provider added?

Accessing the Dify Cloud

• I have problems connecting to "cloud.dify.ai", from China main land directly.
• I can only access it with VPN that enables me accessing Google. I can not understand why.

[NullUserSIe]

EDIT: I successfully added one more specified into the list, finally.

[safiraauliar]

Using self hosted docker (version 1.0.0) face a problem when try to use llamaparse plugin

• In Dify documentation about Dify - llamaparse pluggin, we must set up the api key only from llamaparse cloud and I've successfully set up the api
  

• but when i try create chatflow to parsing pdf file from local upload. it give an error "Request URL is missing" eventho in llamaparse pluggin there's no config to add url except upload file from "paste file link", the full error look like below :
  

• In image below i tried to upload document from "paste file link" method, it give a different error response but still related to URL problem
  

Question

Is it mandatory to add url while setting up the config of llamaparse pluggin? but where can i put the config if it was mandatory?

[safiraauliar]

• from first error I've uploaded the document from local file to a dify server and it was say "Request url is missing". based on the image above the mandatory parameter is transfer_method, type, upload_file_id, and url. but the default local upload config there's no url setting
• In 2nd error I tried to upload file using "paste file link" method from this link https://arxiv.org/pdf/1706.03762 (I've been tried to accessed the url in my browser and it was work) after the document was successfully uploaded from the link, the url change to '/files/cdf1b989-b26f-4cb9-b156-b363e8c9c565/file-preview?timestamp=1741663761&nonce=24c025f2520c4fb38227d857541784a7&sign=vgOvku4iYX3XZiw_L6LfFORJXTeJJfyVqbHa1JkkMVA=" ' based on the error mesage after 3 attempts and caused an error invalid_param.
• I'm using dify in my ubuntu server so the domain is http://<my_ip_address>/

[NullUserSIe]

omitted

What i can think, is Dify get buggy with your file. Have you ever tried to access the url, such ashttp://<my_ip_address>/files/cdf1b989-b26f-4cb9-b156-b363e8c9c565/file-preview? Is that accessable?

[safiraauliar]

when i tried to access the link, it's not provided any file (file not found)

• i saw the headers network said that is successfully upload
  

and this is the response

• I try using other public link to get pdf file still have a same error. I'm so clueless :(

My other question

From the discussion above was focused on "upload file link" method. How about Local upload method, why is it mandatory to provide the url?

[NullUserSIe]

I don't know either, I am not a developer of Dify. That is just an assumption of mine to say, Dify has lost your file after uploaded, and make it inaccessable for the node, with stamp invalid or just no file. You can search the Issues of Dify now, and if really not reported please, i suggest you to fire a ticket for them. Try describe it in details. Maybe just cause the node, but i am afraid it is 1.0.0 broke it.

[safiraauliar]

Okee thanks for your help and advice mate

[pamdla]

遇到问题：比较大的pdf文件解析做Embedding的时候，出现超时被杀错误

api-1            | 2025-03-12 03:40:23.690 WARNING [Dummy-5] [provider_configuration.py:963] - get custom model schema failed, PluginDaemonInternalServerError: killed by timeout

这个如何解决呢？

[nousemail]

After upgrading to versions 1.0.0 and 1.0.1, the local model cannot be saved successfully when added. I saw online that others also encountered the same problem. After returning version 0.15.3, everything was normal.