feat: add context and implement automated subtree synchronization wor…

.github/workflows/sync-subtrees.yml

@@ -0,0 +1,50 @@
+name: Synchronize Remote Repository Context
+
+on:
+  schedule:
+    # Run daily at 00:00 UTC
+    - cron: '0 0 * * *'
+  workflow_dispatch:
+
+concurrency:
+  group: sync-subtrees
+  cancel-in-progress: false
+
+permissions:
+  contents: write
+
+jobs:
+  sync-context:
+    name: Sync repository context
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@11d5960a326750d5838078e36cf38b85af677262 # v4.4.0
+        with:
+          persist-credentials: false
+
+      - name: Set up Python
+        uses: actions/setup-python@a26af69be951a213d495a4c3e4e4022e16d87065 # v5.6.0
+        with:
+          python-version: '3.11'
+
+      - name: Install dependencies
+        run: |
+          python -m pip install --upgrade pip
+          pip install httpx==0.28.1
+
+      - name: Run subtree context sync script
+        run: |
+          python scripts/update_subtrees.py
+
+      - name: Configure git remote authentication
+        env:
+          GH_TOKEN: ${{ github.token }}
+        run: |
+          git config --local http.https://github.com/.extraheader "AUTHORIZATION: basic $(echo -n x-access-token:${GH_TOKEN} | base64)"
+
+      - name: Commit and push updated context files
+        uses: stefanzweifel/git-auto-commit-action@b863ae1933cb653a53c021fe36dbb774e1fb9403 # v5.2.0
+        with:
+          commit_message: "chore(context): auto-sync client repo .agent context files"
+          file_pattern: "repos/**"

context.py

@@ -0,0 +1,78 @@
+import logging
+from pathlib import Path
+
+logger = logging.getLogger("pr-dashboard.context")
+
+MAX_FILE_CHARS = 20_000
+MAX_TOTAL_CHARS = 120_000
+
+
+def get_repo_dir(repo_name: str) -> Path | None:
+    """Find repository context directory inside repos/ or workspace."""
+    bot_root = Path(__file__).resolve().parent
+    candidates = [
+        bot_root / "repos" / repo_name,
+        bot_root / repo_name,
+        bot_root.parent / repo_name,
+    ]
+    for c in candidates:
+        if c.exists() and c.is_dir():
+            return c
+    return None
+
+
+def load_full_repo_context(repo_name: str) -> str:
+    """Dynamically scan and load ALL markdown files inside repos/<repo_name>/ without hardcoding file lists."""
+    repo_dir = get_repo_dir(repo_name)
+    if not repo_dir:
+        logger.warning(f"Repository directory for '{repo_name}' not found.")
+        return f"=== REPOSITORY: {repo_name} ==="
+
+    context_parts = [f"=== REPOSITORY: {repo_name} ==="]
+    loaded_files = 0
+    skipped_files = 0
+    total_chars = 0
+
+    # Recursively scan for all .md files inside the target repository directory
+    for md_file in sorted(repo_dir.rglob("*.md")):
+        # Skip internal cache or git folders
+        if any(part.startswith(".") and part not in [".agent"] for part in md_file.parts):
+            continue
+
+        if total_chars >= MAX_TOTAL_CHARS:
+            skipped_files += 1
+            continue
+
+        rel_path = md_file.relative_to(repo_dir)
+        try:
+            content = md_file.read_text(encoding="utf-8").strip()
+        except Exception as e:
+            logger.error(f"Error reading context file {md_file}: {e}")
+            continue
+
+        if not content:
+            continue
+
+        truncated = len(content) > MAX_FILE_CHARS
+        if truncated:
+            content = content[:MAX_FILE_CHARS]
+
+        remaining = MAX_TOTAL_CHARS - total_chars
+        if len(content) > remaining:
+            content = content[:remaining]
+            truncated = True
+
+        suffix = "\n... [truncated]" if truncated else ""
+        section = f"--- {rel_path} ---\n{content}{suffix}"
+        context_parts.append(section)
+        total_chars += len(section)
+        loaded_files += 1
+
+    if skipped_files:
+        logger.warning(
+            f"Context budget ({MAX_TOTAL_CHARS} chars) reached for '{repo_name}'; "
+            f"skipped {skipped_files} additional markdown file(s)."
+        )
+
+    logger.info(f"Loaded {loaded_files} markdown context files for '{repo_name}' ({total_chars} chars)")
+    return "\n\n".join(context_parts)

main.py

@@ -2,7 +2,7 @@
 main.py — entry point
 
 Flow:
-  1. Load context.md (repo context — what MiniChain is, what's already built)
+  1. Load complete target repo context from repos/<repo_name>/ via context.py
   2. Fetch all PRs + extract CodeRabbit walkthrough + changes only
   3. One combined Ollama call -> groups PRs by problem
   4. Deep Ollama analysis per conflict group (all PRs in group together)
@@ -11,29 +11,19 @@
 
 Run: python main.py
 Requires: gh (authenticated), ollama running on localhost:11434
-Optional: context.md in same folder (drop it in when ready)
+Context Sync: python scripts/update_subtrees.py (syncs target repo context into repos/<repo_name>/)
 """
 
 import os, time, webbrowser
 from github   import fetch_prs, fetch_pr_files, fetch_coderabbit_sections, check_gh_auth, REPO
 from ollama   import check_ollama
 from grouping import resolve_groups
 from render   import build_conflict_html, build_isolated_html
+from context  import load_full_repo_context
 
-OUT_DIR      = os.path.dirname(os.path.abspath(__file__))
-CONTEXT_FILE = os.path.join(OUT_DIR, "context.md")
+OUT_DIR = os.path.dirname(os.path.abspath(__file__))
 
 
-def load_context():
-    if os.path.exists(CONTEXT_FILE):
-        with open(CONTEXT_FILE, "r", encoding="utf-8") as f:
-            content = f.read().strip()
-        print(f"  Loaded context.md ({len(content)} chars)")
-        return content
-    print("  No context.md found — running without repo context")
-    print("  (Drop context.md in the same folder to enable it)")
-    return ""
-
 
 def _build_fallback_pr(raw):
     """Build a minimal PR dict when fetching fails — keeps the pipeline alive."""
@@ -99,8 +89,9 @@ def main():
         print("ERROR: Ollama not reachable at localhost:11434")
         return
 
-    print(f"\nLoading repo context...")
-    repo_context = load_context()
+    target_repo_name = REPO.split("/")[-1]
+    print(f"\nLoading full repo context for '{target_repo_name}'...")
+    repo_context = load_full_repo_context(target_repo_name)
 
     print(f"\nFetching PRs for {REPO}...")
     raw_prs = fetch_prs()

repo_metadata.py

@@ -0,0 +1,14 @@
+REPO_METADATA = {
+    "SocialShareButton": {
+        "url": "https://github.com/kpj2006/SocialShareButton/tree/matt-skills",
+    },
+    "OrgExplorer": {
+        "url": "https://github.com/kpj2006/OrgExplorer/tree/skills",
+    },
+    "GSoC-Proposal-Assistant": {
+        "url": "https://github.com/kpj2006/GSoC-Proposal-Assistant",
+    },
+    "Template-Repo": {
+        "url": "https://github.com/AOSSIE-Org/Template-Repo",
+    },
+}

scripts/update_subtrees.py

@@ -0,0 +1,110 @@
+import os
+import re
+import sys
+import logging
+import httpx
+from pathlib import Path
+
+# Add parent directory to sys.path to import repo_metadata
+script_dir = Path(__file__).resolve().parent
+bot_root = script_dir.parent
+sys.path.insert(0, str(bot_root))
+
+from repo_metadata import REPO_METADATA
+
+logging.basicConfig(level=logging.INFO, format="%(asctime)s [%(levelname)s] %(message)s")
+logger = logging.getLogger("subtree-sync")
+
+# Key context files to sync from remote client repositories
+KNOWN_CONTEXT_FILES = [
+    ".agent/info/operational-data.md",
+    ".agent/core/architecture.md",
+    ".agent/core/code-mapping.md",
+    ".agent/core/edge-cases.md",
+    ".agent/core/examples.md",
+    ".agent/instructions/setup.md",
+    ".agent/instructions/testing.md",
+    ".agent/instructions/deployment.md",
+    ".agent/instructions/ci-cd.md",
+    ".agent/instructions/format-guide.md",
+    ".agent/instructions/checklist.md",
+    ".agent/instructions/bad-patterns.md",
+    "references/checklist.md",
+    "references/format-guide.md",
+    "references/bad-patterns.md",
+    "SKILL.md",
+    "AGENTS.md",
+    "README.md",
+]
+
+
+def parse_github_url(url: str) -> tuple[str, str, str]:
+    """Parse owner, repo, and ref from GitHub URL."""
+    match = re.match(r"https://github\.com/([^/]+)/([^/]+)(?:/tree/([^/]+))?", url)
+    if match:
+        owner, repo, ref = match.groups()
+        repo = repo.removesuffix(".git")
+        ref = ref or "main"
+        return owner, repo, ref
+    return "", "", "main"
+
+
+def sync_repo_context(repo_name: str, meta: dict, client: httpx.Client):
+    """Fetch specified .agent context files directly from raw GitHub endpoint."""
+    url = meta.get("url")
+    if not url:
+        logger.warning(f"No URL defined for {repo_name}, skipping.")
+        return
+
+    owner, repo, ref = parse_github_url(url)
+    if not owner or not repo:
+        logger.error(f"Invalid GitHub URL for {repo_name}: {url}")
+        return
+
+    logger.info(f"Syncing context for '{repo_name}' ({owner}/{repo}@{ref})...")
+    target_dir = bot_root / "repos" / repo_name
+    target_dir.mkdir(parents=True, exist_ok=True)
+
+    headers = {"User-Agent": "PullRequestDashboard-Context-Sync"}
+    downloaded_count = 0
+    removed_count = 0
+
+    for rel_file in KNOWN_CONTEXT_FILES:
+        raw_url = f"https://raw.githubusercontent.com/{owner}/{repo}/{ref}/{rel_file}"
+        dest_path = target_dir / rel_file
+        try:
+            res = client.get(raw_url, headers=headers)
+            if res.status_code == 200:
+                dest_path.parent.mkdir(parents=True, exist_ok=True)
+                dest_path.write_bytes(res.content)
+                logger.info(f"Downloaded: repos/{repo_name}/{rel_file}")
+                downloaded_count += 1
+            elif res.status_code == 404 and dest_path.exists():
+                # Confirmed absent upstream: remove the stale local copy so it
+                # doesn't keep getting fed into the dashboard prompt.
+                dest_path.unlink()
+                logger.info(f"Removed stale file (deleted upstream): repos/{repo_name}/{rel_file}")
+                removed_count += 1
+            # Any other status (rate limit, server error, etc.) is not a
+            # confirmed absence — leave the existing local copy untouched.
+        except Exception as e:
+            logger.debug(f"Failed downloading {rel_file} for {repo_name}: {e}")
+
+    logger.info(
+        f"Synced {downloaded_count} context files into repos/{repo_name} "
+        f"({removed_count} stale file(s) removed)"
+    )
+
+
+def sync_all_subtrees():
+    os.chdir(bot_root)
+    repos_dir = bot_root / "repos"
+    repos_dir.mkdir(exist_ok=True)
+
+    with httpx.Client(timeout=20.0, follow_redirects=True) as client:
+        for repo_name, meta in REPO_METADATA.items():
+            sync_repo_context(repo_name, meta, client)
+
+
+if __name__ == "__main__":
+    sync_all_subtrees()