Skip to content

GERO

GERO #1936

name: Similar Issue Search & Discussion Suggestion
on:
issues:
types: [opened]
pull_request:
types: [opened]
jobs:
analyze:
# Disabling for now due to bugs
if: false
runs-on: ubuntu-latest
permissions:
issues: write
pull-requests: write
steps:
- name: Checkout deja-view
uses: actions/checkout@v6
with:
repository: "bdougie/deja-view"
# Option 1: Pin to specific commit (most secure)
ref: "45e7696" # Pin to specific commit for stability
# Option 2: Use a tag (recommended after creating releases)
# ref: 'v1.0.0'
# Option 3: Use main branch (least secure, gets latest changes)
# ref: 'main'
path: deja-view
- name: Setup Python
uses: actions/setup-python@v6
with:
python-version: "3.11"
- name: Install dependencies
run: |
cd deja-view
pip install -r requirements.txt
- name: Index repository if needed
env:
CHROMA_API_KEY: ${{ secrets.CHROMA_CLOUD_API_KEY }}
CHROMA_TENANT: ${{ secrets.CHROMA_TENANT }}
CHROMA_DATABASE: ${{ secrets.CHROMA_DATABASE }}
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
run: |
cd deja-view
# Index the repository (this will update existing index)
python cli.py index "${{ github.repository }}" --max-issues 200 || true
- name: Search for similar issues
id: search
env:
CHROMA_API_KEY: ${{ secrets.CHROMA_CLOUD_API_KEY }}
CHROMA_TENANT: ${{ secrets.CHROMA_TENANT }}
CHROMA_DATABASE: ${{ secrets.CHROMA_DATABASE }}
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
run: |
cd deja-view
# Determine if this is an issue or PR
if [ "${{ github.event_name }}" = "pull_request" ]; then
ISSUE_URL="https://github.com/${{ github.repository }}/pull/${{ github.event.pull_request.number }}"
THRESHOLD="0.85"
TOP_K="5"
else
ISSUE_URL="https://github.com/${{ github.repository }}/issues/${{ github.event.issue.number }}"
THRESHOLD="0.7"
TOP_K="5"
fi
# Create a Python script to search and format as JSON
cat << 'PYTHON_SCRIPT' > search_similar.py
import sys
import json
import os
from github_similarity_service import SimilarityService
issue_url = sys.argv[1]
threshold = float(sys.argv[2])
top_k = int(sys.argv[3])
# Parse issue number from URL
issue_number = int(issue_url.split('/')[-1])
repo_parts = issue_url.split('/')
owner = repo_parts[3]
repo = repo_parts[4]
service = SimilarityService()
# Find similar issues
similar = service.find_similar_issues(
owner=owner,
repo=repo,
issue_number=issue_number,
top_k=top_k,
min_similarity=threshold
)
# Format as JSON
print(json.dumps(similar))
PYTHON_SCRIPT
# Search for similar issues
SIMILAR_ISSUES=$(python search_similar.py "$ISSUE_URL" "$THRESHOLD" "$TOP_K")
# Save to output
echo "similar_issues<<EOF" >> $GITHUB_OUTPUT
echo "$SIMILAR_ISSUES" >> $GITHUB_OUTPUT
echo "EOF" >> $GITHUB_OUTPUT
# Check if any similar issues found
if [ "$(echo "$SIMILAR_ISSUES" | jq '. | length')" -gt 0 ]; then
echo "has_similar=true" >> $GITHUB_OUTPUT
else
echo "has_similar=false" >> $GITHUB_OUTPUT
fi
- name: Check if issue should be a discussion
id: discussion_check
if: github.event_name == 'issues'
env:
CHROMA_API_KEY: ${{ secrets.CHROMA_CLOUD_API_KEY }}
CHROMA_TENANT: ${{ secrets.CHROMA_TENANT }}
CHROMA_DATABASE: ${{ secrets.CHROMA_DATABASE }}
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
run: |
cd deja-view
# Create a Python script to analyze the issue
cat << 'PYTHON_SCRIPT' > check_discussion.py
import os
import json
import re
from github_similarity_service import SimilarityService
# Get issue details from environment
issue_title = os.environ.get('ISSUE_TITLE', '')
issue_body = os.environ.get('ISSUE_BODY', '')
issue_labels = json.loads(os.environ.get('ISSUE_LABELS', '[]'))
# Mock Issue object for analysis
class MockIssue:
def __init__(self):
self.title = issue_title
self.body = issue_body
self.labels = issue_labels
self.state = 'open'
self.number = int(os.environ.get('ISSUE_NUMBER', 0))
self.created_at = ''
self.updated_at = ''
self.url = ''
service = SimilarityService()
issue = MockIssue()
score, reasons = service._calculate_discussion_score(issue)
result = {
'score': score,
'reasons': reasons,
'should_be_discussion': score >= 0.5,
'confidence': 'high' if score >= 0.7 else 'medium' if score >= 0.5 else 'low'
}
print(json.dumps(result))
PYTHON_SCRIPT
# Run the analysis
DISCUSSION_ANALYSIS=$(ISSUE_TITLE="${{ github.event.issue.title }}" \
ISSUE_BODY='${{ github.event.issue.body }}' \
ISSUE_LABELS='${{ toJson(github.event.issue.labels.*.name) }}' \
ISSUE_NUMBER="${{ github.event.issue.number }}" \
python check_discussion.py)
echo "analysis<<EOF" >> $GITHUB_OUTPUT
echo "$DISCUSSION_ANALYSIS" >> $GITHUB_OUTPUT
echo "EOF" >> $GITHUB_OUTPUT
# Extract whether it should be a discussion
SHOULD_BE_DISCUSSION=$(echo "$DISCUSSION_ANALYSIS" | jq -r '.should_be_discussion')
echo "should_be_discussion=$SHOULD_BE_DISCUSSION" >> $GITHUB_OUTPUT
# Extract confidence level
CONFIDENCE=$(echo "$DISCUSSION_ANALYSIS" | jq -r '.confidence')
echo "confidence=$CONFIDENCE" >> $GITHUB_OUTPUT
- name: Comment with analysis
uses: actions/github-script@v8
with:
script: |
let comment = '';
const hasSimilar = '${{ steps.search.outputs.has_similar }}' === 'true';
const isPR = '${{ github.event_name }}' === 'pull_request';
const shouldBeDiscussion = !isPR && '${{ steps.discussion_check.outputs.should_be_discussion }}' === 'true';
// Add similar issues section if found
if (hasSimilar) {
const similarIssues = JSON.parse(`${{ steps.search.outputs.similar_issues }}`);
if (isPR) {
// Filter only open issues for PRs
const openIssues = similarIssues.filter(issue => issue.state === 'open');
if (openIssues.length > 0) {
// Simple table format for PRs
comment += '## 📋 Related Open Issues\n\n';
comment += 'These open issues appear to be related to this PR (≥85% similarity):\n\n';
comment += '| # | Title | Similarity |\n';
comment += '|---|-------|------------|\n';
openIssues.forEach(issue => {
const similarity = (issue.similarity * 100).toFixed(0);
const title = issue.title.length > 50 ? issue.title.substring(0, 47) + '...' : issue.title;
comment += `| [#${issue.number}](${issue.url}) | ${title} | ${similarity}% |\n`;
});
comment += '\n';
comment += '_Consider closing related issues if this PR addresses them._\n';
} else {
// Don't add any comment if no open issues found
return;
}
} else {
// Original format for issues
comment += '## 🔍 Similar Issues Found\n\n';
comment += 'I found some existing issues that might be related:\n\n';
similarIssues.forEach((issue, index) => {
const similarity = (issue.similarity * 100).toFixed(1);
const state = issue.state === 'open' ? '🟢' : '🔴';
comment += `${index + 1}. ${state} [#${issue.number}: ${issue.title}](${issue.url}) (${similarity}% similar)\n`;
if (issue.summary) {
comment += ` > ${issue.summary.substring(0, 150)}${issue.summary.length > 150 ? '...' : ''}\n`;
}
comment += '\n';
});
comment += '\n';
}
}
// Add discussion suggestion if applicable
if (shouldBeDiscussion) {
const analysis = JSON.parse(`${{ steps.discussion_check.outputs.analysis }}`);
const confidence = '${{ steps.discussion_check.outputs.confidence }}';
if (hasSimilar) {
comment += '---\n\n';
}
comment += '## 💬 Discussion Suggestion\n\n';
const confidenceEmoji = {
'high': '🟢',
'medium': '🟡',
'low': '🔴'
};
comment += `This issue might be better suited as a **GitHub Discussion** ${confidenceEmoji[confidence] || ''}\n\n`;
comment += `**Confidence**: ${confidence.charAt(0).toUpperCase() + confidence.slice(1)} (${(analysis.score * 100).toFixed(0)}%)\n\n`;
if (analysis.reasons && analysis.reasons.length > 0) {
comment += '**Reasons**:\n';
analysis.reasons.forEach(reason => {
comment += `- ${reason}\n`;
});
comment += '\n';
}
comment += 'GitHub Discussions are great for:\n';
comment += '- Questions and help requests\n';
comment += '- Feature requests and ideas\n';
comment += '- General feedback and brainstorming\n';
comment += '- Community conversations\n\n';
comment += '_Consider converting this to a discussion if it\'s not a bug report or actionable task._\n';
}
// Only comment if we have something to say
if (comment) {
comment += '\n---\n';
comment += '_This analysis was automatically generated by [deja-view](https://github.com/bdougie/deja-view)._';
if ('${{ github.event_name }}' === 'pull_request') {
await github.rest.issues.createComment({
owner: context.repo.owner,
repo: context.repo.repo,
issue_number: context.payload.pull_request.number,
body: comment
});
} else {
await github.rest.issues.createComment({
owner: context.repo.owner,
repo: context.repo.repo,
issue_number: context.issue.number,
body: comment
});
}
}
- name: Add labels
if: github.event_name == 'issues'
uses: actions/github-script@v8
with:
script: |
const labels = [];
// Check for duplicate
if ('${{ steps.search.outputs.has_similar }}' === 'true') {
const similarIssues = JSON.parse(`${{ steps.search.outputs.similar_issues }}`);
const hasDuplicate = similarIssues.some(issue => issue.similarity > 0.9);
if (hasDuplicate) {
labels.push('potential-duplicate');
}
}
// Check for discussion suggestion
if ('${{ steps.discussion_check.outputs.should_be_discussion }}' === 'true') {
const confidence = '${{ steps.discussion_check.outputs.confidence }}';
if (confidence === 'high') {
labels.push('should-be-discussion');
} else if (confidence === 'medium') {
labels.push('discussion');
}
}
// Add labels if any
if (labels.length > 0) {
await github.rest.issues.addLabels({
owner: context.repo.owner,
repo: context.repo.repo,
issue_number: context.issue.number,
labels: labels
});
}