refactor: simplify ontology loading, ambiguity, depencencies

.github/workflows/cleanup_images.yaml

@@ -0,0 +1,192 @@
+name: Cleanup Development Images
+
+on:
+  # Clean up when PRs are closed
+  pull_request:
+    types: [closed]
+
+  # Clean up when branches are deleted
+  delete:
+
+  # Scheduled cleanup for old images (runs weekly)
+  schedule:
+    - cron: '0 2 * * 0'  # Every Sunday at 2 AM UTC
+
+  # Manual trigger for cleanup
+  workflow_dispatch:
+    inputs:
+      days_old:
+        description: 'Delete images older than X days'
+        required: false
+        default: '100'
+        type: string
+      tag_pattern:
+        description: 'Tag pattern to clean (e.g., pr-*, develop, feature-*)'
+        required: false
+        default: 'pr-*'
+        type: string
+
+jobs:
+  cleanup-pr-images:
+    # Only run for closed PRs
+    if: github.event_name == 'pull_request' && github.event.action == 'closed'
+    runs-on: ubuntu-latest
+    permissions:
+      packages: write
+
+    steps:
+      - name: Delete PR image
+        uses: actions/github-script@v7
+        with:
+          script: |
+            const prNumber = context.payload.pull_request.number;
+            const packageName = 'git-metadata-extractor';
+            const tag = `pr-${prNumber}`;
+
+            try {
+              // Get package version for the PR tag
+              const { data: versions } = await github.rest.packages.getAllPackageVersionsForPackageOwnedByOrg({
+                package_type: 'container',
+                package_name: packageName,
+                org: context.repo.owner
+              });
+
+              const prVersion = versions.find(v => 
+                v.metadata && v.metadata.container && 
+                v.metadata.container.tags.includes(tag)
+              );
+
+              if (prVersion) {
+                await github.rest.packages.deletePackageVersionForOrg({
+                  package_type: 'container',
+                  package_name: packageName,
+                  org: context.repo.owner,
+                  package_version_id: prVersion.id
+                });
+                console.log(`✅ Deleted image with tag: ${tag}`);
+              } else {
+                console.log(`ℹ️ No image found with tag: ${tag}`);
+              }
+            } catch (error) {
+              console.log(`⚠️ Error cleaning up image ${tag}:`, error.message);
+              // Don't fail the workflow if cleanup fails
+            }
+
+  cleanup-branch-images:
+    # Only run when branches are deleted
+    if: github.event_name == 'delete' && github.event.ref_type == 'branch'
+    runs-on: ubuntu-latest
+    permissions:
+      packages: write
+
+    steps:
+      - name: Delete branch image
+        uses: actions/github-script@v7
+        with:
+          script: |
+            const branchName = context.payload.ref;
+            const packageName = 'git-metadata-extractor';
+            // Sanitize branch name to match Docker tag format
+            const tag = branchName.replace(/[^a-zA-Z0-9._-]/g, '-').toLowerCase();
+
+            try {
+              const { data: versions } = await github.rest.packages.getAllPackageVersionsForPackageOwnedByOrg({
+                package_type: 'container',
+                package_name: packageName,
+                org: context.repo.owner
+              });
+
+              const branchVersion = versions.find(v => 
+                v.metadata && v.metadata.container && 
+                v.metadata.container.tags.includes(tag)
+              );
+
+              if (branchVersion) {
+                await github.rest.packages.deletePackageVersionForOrg({
+                  package_type: 'container',
+                  package_name: packageName,
+                  org: context.repo.owner,
+                  package_version_id: branchVersion.id
+                });
+                console.log(`✅ Deleted image with tag: ${tag} (branch: ${branchName})`);
+              } else {
+                console.log(`ℹ️ No image found with tag: ${tag} (branch: ${branchName})`);
+              }
+            } catch (error) {
+              console.log(`⚠️ Error cleaning up image ${tag}:`, error.message);
+            }
+
+  cleanup-old-images:
+    # Run on schedule or manual trigger
+    if: github.event_name == 'schedule' || github.event_name == 'workflow_dispatch'
+    runs-on: ubuntu-latest
+    permissions:
+      packages: write
+
+    steps:
+      - name: Cleanup old development images
+        uses: actions/github-script@v7
+        with:
+          script: |
+            const packageName = 'git-metadata-extractor';
+            const daysOld = parseInt('${{ inputs.days_old || 100 }}');
+            const tagPattern = '${{ inputs.tag_pattern }}' || 'pr-*';
+            const cutoffDate = new Date();
+            cutoffDate.setDate(cutoffDate.getDate() - daysOld);
+
+            console.log(`🧹 Cleaning up images older than ${daysOld} days (before ${cutoffDate.toISOString()})`);
+            console.log(`🎯 Tag pattern: ${tagPattern}`);
+
+            try {
+              const { data: versions } = await github.rest.packages.getAllPackageVersionsForPackageOwnedByOrg({
+                package_type: 'container',
+                package_name: packageName,
+                org: context.repo.owner,
+                per_page: 100
+              });
+
+              let deletedCount = 0;
+
+              for (const version of versions) {
+                const createdAt = new Date(version.created_at);
+                const tags = version.metadata?.container?.tags || [];
+
+                // Skip if not old enough
+                if (createdAt > cutoffDate) continue;
+
+                // Check if any tag matches the pattern
+                const matchesPattern = tags.some(tag => {
+                  if (tagPattern === 'pr-*') return tag.startsWith('pr-');
+                  if (tagPattern === 'feature-*') return tag.startsWith('feature-');
+                  if (tagPattern === 'develop') return tag === 'develop';
+                  return tag.includes(tagPattern.replace('*', ''));
+                });
+
+                // Skip protected tags (latest, version numbers)
+                const hasProtectedTag = tags.some(tag => 
+                  tag === 'latest' || 
+                  /^\d+\.\d+\.\d+/.test(tag)
+                );
+
+                if (matchesPattern && !hasProtectedTag) {
+                  try {
+                    await github.rest.packages.deletePackageVersionForOrg({
+                      package_type: 'container',
+                      package_name: packageName,
+                      org: context.repo.owner,
+                      package_version_id: version.id
+                    });
+                    console.log(`✅ Deleted old image: ${tags.join(', ')} (created: ${createdAt.toISOString()})`);
+                    deletedCount++;
+                  } catch (error) {
+                    console.log(`⚠️ Error deleting image ${tags.join(', ')}:`, error.message);
+                  }
+                }
+              }
+
+              console.log(`🎉 Cleanup completed! Deleted ${deletedCount} images.`);
+
+            } catch (error) {
+              console.error('❌ Error during cleanup:', error.message);
+              throw error;
+            }

.github/workflows/publish_image_in_GHCR.yaml

@@ -0,0 +1,83 @@
+name: Build and Publish Docker Images
+
+on:
+  push:
+    branches: [ "main", "develop" ]
+  pull_request:
+    branches: [ "main", "develop" ]
+
+  workflow_dispatch:
+
+jobs:
+  build-and-publish:
+    runs-on: ubuntu-latest
+    permissions:
+      contents: write # needed to create the release
+      packages: write # needed to publish the image
+
+    # Skip building images for draft PRs or very old branches
+    if: |
+      github.event_name == 'push' || 
+      github.event_name == 'workflow_dispatch' ||
+      (github.event_name == 'pull_request' && 
+       github.event.pull_request.draft == false)
+
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v4
+
+      - name: Extract version from pyproject.toml
+        id: project_version
+        run: |
+          VERSION=$(sed -n '/^\[project\]/,/^\[/p' pyproject.toml | grep '^version =' | sed -E 's/version = "([^"]+)"/\1/')
+          echo "version=${VERSION}" >> $GITHUB_OUTPUT
+
+      - name: Log in to GitHub Container Registry
+        uses: docker/login-action@v3
+        with:
+          registry: ghcr.io
+          username: ${{ github.actor }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Extract metadata for Docker
+        id: meta
+        uses: docker/metadata-action@v5
+        with:
+          images: ghcr.io/${{ github.repository }}
+          tags: |
+            # For main branch: latest and version tags
+            type=raw,value=latest,enable={{is_default_branch}}
+            type=raw,value=${{ steps.project_version.outputs.version }},enable={{is_default_branch}}
+            # For develop branch: develop tag
+            type=raw,value=develop,enable=${{ github.ref == 'refs/heads/develop' }}
+            # For PRs: pr-{number} tag
+            type=ref,event=pr,prefix=pr-
+            # For feature branches: branch name (sanitized)
+            type=ref,event=branch,enable=${{ github.ref != 'refs/heads/main' && github.ref != 'refs/heads/develop' }}
+          labels: |
+            org.opencontainers.image.title=${{ github.repository }}
+            org.opencontainers.image.description=${{ github.event.repository.description }}
+            org.opencontainers.image.url=${{ github.event.repository.html_url }}
+            org.opencontainers.image.source=${{ github.event.repository.clone_url }}
+            org.opencontainers.image.revision=${{ github.sha }}
+            org.opencontainers.image.licenses=${{ github.event.repository.license.spdx_id }}
+
+      - name: Build and push Docker image
+        uses: docker/build-push-action@v5
+        with:
+          context: .
+          file: tools/image/Dockerfile
+          platforms: linux/amd64
+          push: true
+          tags: ${{ steps.meta.outputs.tags }}
+          labels: ${{ steps.meta.outputs.labels }}
+
+      - name: Create GitHub Release
+        # Only create releases for main branch pushes
+        if: github.ref == 'refs/heads/main' && github.event_name == 'push'
+        uses: softprops/action-gh-release@v2
+        with:
+          tag_name: v${{ steps.project_version.outputs.version }}
+          name: Release v${{ steps.project_version.outputs.version }}
+          body_path: CHANGELOG.md
+          fail_on_unmatched_files: true

pyproject.toml

@@ -5,16 +5,20 @@ authors = [
   { name = "Martin Fontanet", email = "martin.fontanet@epfl.ch" },
 ]
 description = "This a repository turns strings into things"
-readme = "README.md"
-requires-python = ">=3.13"
+requires-python = ">=3.11"
 version = "0.0.1"
 
 dependencies = [
     "dotenv>=0.9.9",
+    "fastapi>=0.116.1",
     "openai-agents>=0.0.8",
     "pydantic>=2.10.6",
+    "pydantic-settings>=2.8.1",
     "pyfuzon>=0.3.0",
+    "pytest>=8.4.1",
+    "python-multipart>=0.0.20",
     "rdflib>=7.1.4",
+    "SPARQLWrapper>=2.0.0",
 ]
 [dependency-groups]
 dev = [

src/strings2things/app/__init__.py

@@ -0,0 +1,2 @@
+# app/__init__.py
+# Initialize app module

src/strings2things/app/api/__init__.py

@@ -0,0 +1,2 @@
+# app/api/__init__.py
+# Initialize API module

src/strings2things/app/api/endpoints.py

@@ -0,0 +1,43 @@
+# app/api/endpoints.py
+
+from fastapi import APIRouter, UploadFile, File, Form
+from fastapi.responses import Response
+from rdflib import Graph
+from src.strings2things.app.core.rdf_transformer import RDFTransformer
+from src.strings2things.app.core.ontology_manager import (
+    OntologyManager,
+)  # Assume this exists
+from src.strings2things.app.utils.rdf_utils import (
+    parse_rdf,
+    serialize_rdf,
+)  # Also assume or create
+import logging
+
+router = APIRouter()
+
+ontology_manager = OntologyManager()
+ontology_manager.load_ontologies()
+
+transformer = RDFTransformer(ontology_manager.get_label_map())
+
+
+@router.post("/transform")
+async def transform_rdf(
+    file: UploadFile = File(...), serialization: str = Form("turtle")
+) -> Response:
+    """
+    Accepts an RDF file upload, transforms it using the label map,
+    and returns the modified RDF graph in the requested format.
+    """
+    content = await file.read()
+
+    try:
+        input_graph = parse_rdf(content)
+        transformed_graph = transformer.transform(input_graph)
+        serialized = serialize_rdf(transformed_graph, output_format=serialization)
+
+        return Response(content=serialized, media_type="text/plain")
+
+    except Exception as e:
+        logging.exception("Transformation failed")
+        return Response(content=str(e), status_code=400)

src/strings2things/app/config.py

@@ -0,0 +1,26 @@
+from typing import List
+from pydantic import PrivateAttr
+from pydantic_settings import BaseSettings, SettingsConfigDict
+
+
+class Settings(BaseSettings):
+    GRAPHDB_USERNAME: str
+    GRAPHDB_PASSWORD: str
+    ONTOLOGY_SPARQL_ENDPOINT: str
+    ONTOLOGY_GRAPH_IRIS: str  # raw string from .env
+    FAIL_ON_AMBIGUOUS_LABELS: bool = True
+
+    _graph_iris: List[str] = PrivateAttr()
+
+    def __init__(self, **kwargs):
+        super().__init__(**kwargs)
+        self._graph_iris = [i.strip() for i in self.ONTOLOGY_GRAPH_IRIS.split(",")]
+
+    def get_graph_iris(self) -> List[str]:
+        return self._graph_iris
+
+    model_config = SettingsConfigDict(
+        env_file=".env",
+        env_file_encoding="utf-8",
+        extra="ignore",
+    )

src/strings2things/app/core/__init__.py

@@ -0,0 +1,2 @@
+# app/core/__init__.py
+# Initialize core logic module