Error: HttpError: Not Found

[PsiACE]

Summary

I tried to trigger this workflow through workflow_dispatch, but encountered an error at the step of gpt-translate: Error: HttpError: Not Found. I am not sure about the context, so I hope someone can help me.

Additional information

name: GPT Translate by Input

on:
  workflow_dispatch:
    inputs:
      path:
        description: 'The path to the directory containing the files to translate (e.g., docs/sql-reference/)'
        required: true
        default: 'sql-reference'

permissions:
  id-token: write
  pull-requests: write
  checks: write
  statuses: write
  contents: write

jobs:
  gpt_translate:
    runs-on: ubuntu-latest

    steps:
    - name: Checkout
      uses: actions/checkout@v4
      with:
        fetch-depth: 0
    
    - name: Delete all files in specified cn directory
      run: |
        TRANSLATION_DIR="docs/${{ github.event.inputs.path }}"
        CN_DIR="docs/cn/${{ github.event.inputs.path }}"
        echo "Deleting all files in $CN_DIR"
        rm -rf $CN_DIR
        echo "All files in $CN_DIR have been deleted."
        git add .
    - name: Get all files in specified directory
      id: all_files
      run: |
        TRANSLATION_DIR="docs/${{ github.event.inputs.path }}"
        echo "files=$(find $TRANSLATION_DIR -type f -name '*.md' | sed -e 's/^/.\//' | tr '\n' ' ')" >> $GITHUB_OUTPUT
    - name: Run GPT Translate
      uses: 3ru/gpt-translate@master
      with:
        apikey: ${{ secrets.OPENAI_API_KEY }}
        model: "gpt-4-turbo-preview"
        inputFiles: '${{ steps.all_files.outputs.files }}'
        outputFiles: 'docs/cn/**/*.md'
        languages: 'Simplified-Chinese'
        prompt: "You are a translation engine that has knowledge of databases and is familiar with SQL, HTML and Markdown syntax. \ Databend is a cloud-native data warehouse and an alternative to Snowflake. \ I am translating the Databend documentation for helping users.\ Translate the Markdown content I'll paste later into {targetLanguage}.\ You must strictly follow the rules below.\ - Never change the Markdown markup structure. Don't add or remove links. Do not change any URL.\ - Never change the contents of code blocks even if they appear to have a bug.\ - Always preserve the original line breaks. Do not add or remove blank lines.\ - Never touch the permalink such as `{/*examples*/}` at the end of each heading.\ - Never touch HTML-like tags such as `<Notes>`.\ - Correctly format the document for best rendering."
        token: ${{ secrets.GITHUB_TOKEN }}

Example

https://github.com/datafuselabs/databend-docs/actions/runs/7690600375/job/20954644583

[3ru]

@PsiACE
It was caused by not supporting workflow_dispatch, now that it is supported in v1.1.9, could you please run and try again?

[PsiACE]

@PsiACE It was caused by not supporting workflow_dispatch, now that it is supported in v1.1.9, could you please run and try again?

thanks @3ru , It can run normally now, but the output files is not quite as expected. I would like to ask for your help in checking how to write it correctly.

I hope ./docs/sql-reference/10-sql-commands/50-administration-cmds/show-metrics.md can be correctly mapped to ./docs/cn/sql-reference/10-sql-commands/50-administration-cmds/show-metrics.md. Instead of ./docs/cn/10-sql-commands/50-administration-cmds/show-metrics.md

• https://github.com/datafuselabs/databend-docs/actions/runs/7695506692
• 🌐 Add LLM Translations databendlabs/databend-docs#459

[3ru]

@PsiACE
This is the expected behavior for this bot. My bot replaces the paths it is given up to the same hierarchy.
For example, if I specify output as outputFiles: docs/cn/**/*.md, the top two(Because two levels of docs and cn are specified here) hierarchies up to /**/ are replaced.

./ docs/sql-reference/	10-sql-commands/50-administration-cmds/show-metrics.md
. /docs/sql-reference -> docs/cn/	10-sql-commands/50-administration-cmds/show-metrics.md

To achieve your requirements, you need to do like this

 outputFiles: 'docs/cn/${{ github.event.inputs.path }}/**/*.md'

[PsiACE]

@PsiACE This is the expected behavior for this bot. My bot replaces the paths it is given up to the same hierarchy. For example, if I specify output as outputFiles: docs/cn/**/*.md, the top two(Because two levels of docs and cn are specified here) hierarchies up to /**/ are replaced

I noticed this PR, it still lacks 10-sql-commands. So it still appears to be incorrect.

3ru/databend-docs#2

[3ru]

Certainly, I will check.

PS
docs/cn/${{ github.event.inputs.path }}/**/*.md
This replaced the top three hierarchies, which was not ideal.
I will consider countermeasures.

[3ru]

@PsiACE
I'm sorry, but I couldn't come up with an immediate solution that perfectly meets your ideal. The reason is that I am expecting to change /docs/en to /docs/cn.

As a workaround, I would suggest either moving the input file structure to en before the translation job, or inserting a shell script to modify the output to the desired directory structure after completion.

[PsiACE]

I now realize this problem, and I will reorganize the directory structure to meet the requirements as much as possible. Thank you very much for your answer, and also thank you for this useful action.

If you have a plan to support the terminology reference table, this action will become a real game changer. I've been using gpt-4-turbo-preview recently and it works well, but there are some words that cannot be processed correctly. We need a good way to intervene in this process.

[3ru]

I can empathize with the imperfections of GPT translations.
If possible, I'd appreciate it if you could start a new discussion as a feature request regarding what a “terminology reference table” might entail. As of now, I'm not clear on what it is.