fix: correct licensee version, JSON parsing, and null type cast IN-1105

Signed-off-by: Gašper Grom <gasper.grom@gmail.com>

Author: gaspergrom

scripts/services/docker/Dockerfile.git_integration

@@ -84,11 +84,15 @@ RUN apt-get update && apt-get install -y \
     git \
     ripgrep \
     ruby \
+    libgit2-1.1 \
     ruby-dev \
     build-essential \
+    libgit2-dev \
+    cmake \
+    pkg-config \
     --no-install-recommends \
-    && gem install licensee -v '10.0.0' --no-document \
-    && apt-get remove --autoremove -y ruby-dev build-essential \
+    && gem install licensee -v '9.15.3' --no-document \
+    && apt-get remove --autoremove -y ruby-dev build-essential libgit2-dev cmake pkg-config \
     && rm -rf /var/lib/apt/lists/*
 
 ENV PYTHONUNBUFFERED=1 \

services/apps/git_integration/src/crowdgit/database/crud.py

@@ -286,10 +286,10 @@ async def update_last_processed_commit(repo_id: str, commit_hash: str, branch: s
 async def update_repository_license(repository_id: str, license_spdx: str | None) -> None:
     sql_query = """
     UPDATE public.repositories
-        SET license = $1,
+        SET license = $1::varchar,
         "updatedAt" = NOW()
     WHERE id = $2
-      AND ($1 IS NOT NULL OR license IS NULL)
+      AND ($1::varchar IS NOT NULL OR license IS NULL)
     """
     await execute(sql_query, (license_spdx, repository_id))
 

services/apps/git_integration/src/crowdgit/services/license/license_service.py

@@ -27,9 +27,10 @@ async def detect(self, repo_path: str) -> str | None:
 
         try:
             data = json.loads(output)
-            matched = data.get("matched_license") or {}
-            spdx_id = matched.get("spdx_id")
-            confidence = matched.get("confidence")
+            licenses = data.get("licenses") or []
+            matched_files = data.get("matched_files") or []
+            spdx_id = licenses[0].get("spdx_id") if licenses else None
+            confidence = (matched_files[0].get("matcher") or {}).get("confidence") if matched_files else None
             if spdx_id:
                 self.logger.info(f"License detected: {spdx_id} (confidence={confidence}) in {repo_path}")
             else: