Merge pull request #2 from rsb-23/develop

PR comment token fix

.github/workflows/test.yml

@@ -12,9 +12,9 @@ jobs:
           fetch-depth: 0
 
       - name: Run script
-        id: spelling
         uses: ./
         with:
+          token: ${{ secrets.PAT }}
           files: |
             tests/file.txt
             tests/file.json

README.md

@@ -0,0 +1,16 @@
+# actions/check-typo
+
+This Github Action uses AI to find typos and grammatical errors in PR changes.
+
+## Usage
+Refer [test.yml](./.github/workflows/test.yml)
+```yaml
+  - name: Check typos in data files
+    uses: actions/check-diff-typo@v1
+    with:
+      token: ${{ secrets.PAT }}
+      files: |
+        tests/file.txt
+        tests/file.json
+      GROQ_API_KEY: ${{ secrets.GROQ_API_KEY }}
+```

action.yml

@@ -1,11 +1,15 @@
-name: 'Check-Diff-Typo'
-description: 'Analyzes and comments on PR diff typos for specified files'
+name: 'Check-Typo'
+description: 'Checks diff of specified files for typos and suggests fixes on PR.'
 author: 'Rishabh B'
 branding:
   icon: 'git-pull-request'
   color: 'green'
 
 inputs:
+  token:
+    description: 'PAT with `pull_request:write` permission'
+    required: true
+    default: ${{ github.token }}
   files:
     description: 'Comma-separated list of files to analyze'
     required: true
@@ -14,10 +18,6 @@ inputs:
     description: 'Groq API Key'
     required: true
 
-outputs:
-  comment:
-    description: 'Generated comment markdown'
-    value: ${{ steps.spelling.outputs.comment }}
 
 runs:
   using: "composite"
@@ -28,6 +28,8 @@ runs:
         cache: 'pip'
     - name: Setup python env
       run: |
+        python -m venv .venv
+        source .venv/bin/activate
         python -m pip install --upgrade pip
         pip install -r requirements.txt
       shell: bash
@@ -40,13 +42,11 @@ runs:
       id: spelling
       shell: bash
       env:
-        PR_BASE: ${{ github.event.pull_request.base.ref }}
+        PR_BASE: ${{ github.base_ref }}
         INPUT_FILES: ${{ inputs.files }}
         GROQ_API_KEY: ${{ inputs.GROQ_API_KEY }}
-      run: python src/spell_check.py
-
-    - name: Comment PR
-      uses: peter-evans/create-or-update-comment@v4
-      with:
-        issue-number: ${{ github.event.pull_request.number }}
-        body: ${{ steps.spelling.outputs.comment }}
+        token: ${{ inputs.token }}
+        PR_NO: ${{ github.event.pull_request.number }}
+      run: |
+        source .venv/bin/activate
+        python main.py

local_test.py

@@ -1,18 +1,8 @@
-import os
-import tempfile
-
-from src.spell_check import main
-
-
-def run():
-    if "GITHUB_OUTPUT" not in os.environ:
-        from dotenv import load_dotenv
-
-        load_dotenv()
-        with tempfile.NamedTemporaryFile(mode="w+", delete_on_close=False) as f:
-            os.environ["GITHUB_OUTPUT"] = f.name
-            main()
+from dotenv import load_dotenv
 
+load_dotenv()
 
 if __name__ == "__main__":
-    run()
+    from src.spell_check import main
+
+    main()

main.py

@@ -0,0 +1,4 @@
+from src.spell_check import main
+
+if __name__ == "__main__":
+    main()

requirements.txt

@@ -1 +1,2 @@
 groq==0.13.0
+requests~=2.32.3

src/config.py

@@ -0,0 +1,14 @@
+import os
+
+
+def get_env(x):
+    return os.environ[x]
+
+
+GROQ_API_KEY = get_env("GROQ_API_KEY")
+PAT_TOKEN = get_env("token")
+
+REPO = get_env("GITHUB_REPOSITORY")
+PR_BASE = get_env("PR_BASE")
+PR_NO = get_env("PR_NO")
+INPUT_FILES = [*map(str.strip, get_env("INPUT_FILES").splitlines())]

src/groq_ai.py

@@ -1,15 +1,16 @@
 import json
-import os
 from functools import cache
 
 from groq import NOT_GIVEN, BadRequestError, Groq
 
+from src.config import GROQ_API_KEY
+
 model = ["llama-3.1-8b-instant", "llama3-70b-8192"][1]
 
 
 @cache
 def get_groq_client():
-    return Groq(api_key=os.environ.get("GROQ_API_KEY"))
+    return Groq(api_key=GROQ_API_KEY)
 
 
 def ask_groq(query_text: str, system_content: str, json_schema: dict = None) -> dict:
@@ -25,16 +26,19 @@ def ask_groq(query_text: str, system_content: str, json_schema: dict = None) ->
             response_format=response_format,
         )
         json_str = chat_completion.choices[0].message.content
+        json_str = json_str.strip()
     except BadRequestError as e:
         failed_json = e.body["error"]["failed_generation"]  # noqa
         json_str = failed_json.replace('"""', '"').replace("\n", "\\n")
     if not json_schema:
         return json_str
     try:
-        return json.loads(json_str.replace("\n", ""))
+        # json_str = re.sub("\n +", "", json_str)
+        return json.loads(json_str)
     except json.JSONDecodeError as e:
-        print(e.doc, e.pos)
-        raise
+        print(e.doc.encode(), e.pos)
+        print(f"ERROR : {e.args}")
+        return {"suggestions": ["..."]}
 
 
 def find_typos(query_text):

src/spell_check.py

@@ -1,35 +1,43 @@
-import json
-import os
 import subprocess
 
-from groq_ai import find_typos
+import requests
 
+from src.config import INPUT_FILES, PAT_TOKEN, PR_BASE, PR_NO, REPO
+from src.groq_ai import find_typos
 
-def process_diff(file_path, base_branch):
+
+def process_diff(file_path, base_branch=PR_BASE):
     try:
         # Get diff from PR
         diff_command = f"git diff -U0 origin/{base_branch}... -- {file_path}"
         diff_output = subprocess.check_output(diff_command.split()).decode("utf-8")
     except subprocess.CalledProcessError:
         print("Error in process diff")
-        return 0
+        return -1
 
     # only checks for typos in added text
     new_text = "".join(x for x in diff_output.splitlines(keepends=True) if x.startswith("+"))
     return find_typos(new_text) if new_text else -1
 
 
-def main():
-    # Get required inputs
-    pr_base = os.environ["PR_BASE"]
-    files = [*map(str.strip, os.environ["INPUT_FILES"].splitlines())]
-    print(pr_base, *files, sep="\n")
+def post_comment(comment):
+    url = f"https://api.github.com/repos/{REPO}/issues/{PR_NO}/comments"
+    headers = {
+        "Accept": "application/vnd.github+json",
+        "Authorization": f"Bearer {PAT_TOKEN}",
+        "X-GitHub-Api-Version": "2022-11-28",
+    }
+    resp = requests.post(url, headers=headers, json={"body": comment})
+    resp.raise_for_status()
 
-    results = {file_path: process_diff(file_path, pr_base) for file_path in files if file_path}
+
+def main():
+    # Process diff(s) for each file
+    results = {file_path: process_diff(file_path) for file_path in INPUT_FILES if file_path}
 
     # Create markdown comment for fixes
     flag = False
-    comment = "## Suggested Typo Fixes\n\n"
+    comment = "### Suggested Typo Fixes\n"
     for file_path, suggestion in results.items():
         if suggestion == -1:
             continue
@@ -41,10 +49,8 @@ def main():
     else:
         comment = "### No typos found"
 
-    # Set output for GitHub Actions
-    with open(os.environ["GITHUB_OUTPUT"], "a") as f:
-        f.write(f"comment={json.dumps(comment)}\n")
     print(comment)
+    post_comment(comment)
 
 
 if __name__ == "__main__":

tests/file.json

@@ -1 +1 @@
-{"fruit1": "pineaple", "friut2": "gauva"}
+{"fruit1": "pineaple", "friut2": "graep"}