adjusted design and added miising projects functionality

Author: Tobias Geilen

backend/app/repositories/project_repository.py

@@ -104,4 +104,3 @@ def remove_paper(db: Session, project_id: int, paper_id: int, user_id: int) -> P
         db.refresh(project)
         return project
 
-

backend/app/routes/project_routes.py

@@ -129,4 +129,3 @@ def remove_paper_from_project(
     return ProjectService.remove_paper_from_project(
         db, current_username, project_id, paper_id
     )
-

backend/app/routes/search_routes.py

@@ -1,5 +1,7 @@
-from fastapi import APIRouter, status
+from fastapi import APIRouter, Depends, status
+from sqlalchemy.orm import Session
 
+from app.core.database import get_db
 from app.schemas.search_dto import SearchRequest, SearchResponse
 from app.services.search_service import SearchService
 
@@ -12,10 +14,12 @@
     status_code=status.HTTP_200_OK,
     summary="Search for papers",
 )
-def search(request: SearchRequest) -> SearchResponse:
+def search(request: SearchRequest, db: Session = Depends(get_db)) -> SearchResponse:
     """
-    Returns a list of papers that match the search query
+    Returns a list of papers that match the search query.
+
+    Currently, this returns the 5 most recently fetched papers, regardless
+    of the query string.
     """
 
-    papers = SearchService.search_papers(request.query)
-    return SearchResponse.model_validate(papers)
+    return SearchService.search_papers(db, request.query)

backend/app/schemas/search_dto.py

@@ -10,7 +10,7 @@ class Paper(BaseModel):
     """
 
     paper_id: int
-    doi: str
+    doi: Optional[str]  # DOI is not always reliable or present in the source data
     source: str
     paper_type: str
     title: str

backend/app/services/project_service.py

@@ -49,7 +49,36 @@ def get_project_with_papers(
                 detail="Project not found.",
             )
 
-        papers = [PaperSummary.model_validate(p) for p in project.papers]
+        papers: list[PaperSummary] = []
+        for paper in project.papers:
+            authors_value = None
+            if isinstance(paper.authors, dict):
+                authors_value = paper.authors
+            elif isinstance(paper.authors, list):
+                formatted = []
+                for item in paper.authors:
+                    if isinstance(item, (list, tuple)) and item:
+                        last = item[0] or ""
+                        first = item[1] if len(item) > 1 else ""
+                        middle = item[2] if len(item) > 2 else ""
+                        name = " ".join(part for part in [first, middle, last] if part)
+                        if name:
+                            formatted.append(name)
+                if formatted:
+                    authors_value = {
+                        str(idx): name for idx, name in enumerate(formatted)
+                    }
+            papers.append(
+                PaperSummary(
+                    paper_id=paper.paper_id,
+                    title=paper.title,
+                    authors=authors_value,
+                    abstract=paper.abstract,
+                    published_at=paper.published_at,
+                    url=paper.paper_id_external,
+                    pdf_url=None,
+                )
+            )
         return ProjectWithPapersResponse(
             project=ProjectResponse.model_validate(project),
             papers=papers,
@@ -111,7 +140,36 @@ def add_paper_to_project(
                 detail=str(exc),
             ) from exc
 
-        papers = [PaperSummary.model_validate(p) for p in project.papers]
+        papers: list[PaperSummary] = []
+        for paper in project.papers:
+            authors_value = None
+            if isinstance(paper.authors, dict):
+                authors_value = paper.authors
+            elif isinstance(paper.authors, list):
+                formatted = []
+                for item in paper.authors:
+                    if isinstance(item, (list, tuple)) and item:
+                        last = item[0] or ""
+                        first = item[1] if len(item) > 1 else ""
+                        middle = item[2] if len(item) > 2 else ""
+                        name = " ".join(part for part in [first, middle, last] if part)
+                        if name:
+                            formatted.append(name)
+                if formatted:
+                    authors_value = {
+                        str(idx): name for idx, name in enumerate(formatted)
+                    }
+            papers.append(
+                PaperSummary(
+                    paper_id=paper.paper_id,
+                    title=paper.title,
+                    authors=authors_value,
+                    abstract=paper.abstract,
+                    published_at=paper.published_at,
+                    url=paper.paper_id_external,
+                    pdf_url=None,
+                )
+            )
         return ProjectWithPapersResponse(
             project=ProjectResponse.model_validate(project),
             papers=papers,
@@ -133,10 +191,38 @@ def remove_paper_from_project(
                 detail=str(exc),
             ) from exc
 
-        papers = [PaperSummary.model_validate(p) for p in project.papers]
+        papers: list[PaperSummary] = []
+        for paper in project.papers:
+            authors_value = None
+            if isinstance(paper.authors, dict):
+                authors_value = paper.authors
+            elif isinstance(paper.authors, list):
+                formatted = []
+                for item in paper.authors:
+                    if isinstance(item, (list, tuple)) and item:
+                        last = item[0] or ""
+                        first = item[1] if len(item) > 1 else ""
+                        middle = item[2] if len(item) > 2 else ""
+                        name = " ".join(part for part in [first, middle, last] if part)
+                        if name:
+                            formatted.append(name)
+                if formatted:
+                    authors_value = {
+                        str(idx): name for idx, name in enumerate(formatted)
+                    }
+            papers.append(
+                PaperSummary(
+                    paper_id=paper.paper_id,
+                    title=paper.title,
+                    authors=authors_value,
+                    abstract=paper.abstract,
+                    published_at=paper.published_at,
+                    url=paper.paper_id_external,
+                    pdf_url=None,
+                )
+            )
         return ProjectWithPapersResponse(
             project=ProjectResponse.model_validate(project),
             papers=papers,
         )
 
-

backend/app/services/search_service.py

@@ -1,44 +1,100 @@
 import datetime
 
-from app.core.deps import get_openai_provider, get_specter2_query_embedder
-from app.schemas.search_dto import Paper, SearchResponse
+from sqlalchemy import desc
+from sqlalchemy.orm import Session
+
+from app.models.paper import Paper as PaperModel
+from app.schemas.search_dto import Paper as PaperSchema, SearchResponse
+
+# from app.core.deps import get_openai_provider, get_specter2_query_embedder
 
 
 class SearchService:
     """
-    Service for managing search requests
+    Service for managing search requests.
+
+    NOTE:
+    -----
+    For now, we just return the 5 most recently fetched papers from the
+    database (ordered by `fetched_at DESC`) to establish the connection
+    between the API and the frontend.
+
+    To restore or extend the original behavior that uses OpenAI + SPECTER2
+    embeddings, uncomment and adapt the code in `search_papers` below.
     """
 
     @staticmethod
-    def search_papers(query: str) -> SearchResponse:
+    def search_papers(db: Session, query: str) -> SearchResponse:
         """
-        Returns a list of matching papers based on the query
+        Return the 5 most recent papers from the database.
+
+        The `query` parameter is currently ignored; it's kept for future
+        use when more advanced search logic is implemented.
         """
-        openai_provider = get_openai_provider()
-        keywords = openai_provider.extract_keywords(query)
-
-        embedder = get_specter2_query_embedder()
-        embeddings = embedder.embed_batch(keywords)
-
-        # Additionally embedd the original query -> could also be used for searching
-        embeddings.append(embedder.embed_one(query))
-
-        # Call search module to find matching papers in vector db
-        # Needs to be implemented in the future
-
-        # Mock paper as we cannot implement search as of now
-        example_paper = Paper(
-            paper_id=1,
-            doi="some doi",
-            source="arXiv",
-            paper_type="conference paper",
-            title="Deep Reinforcement Learning for Multi-Agent Voting with Approval Preferences",
-            authors={"main": "Usain Bolt"},
-            abstract="Very good paper about DRL and social choice.",
-            published_at=datetime.date.today(),
-            pdf_url="some url",
-            url="some url",
-            fetched_at=datetime.datetime.now(),
+
+        # --- Original implementation (kept for easy reactivation) ---
+        # openai_provider = get_openai_provider()
+        # keywords = openai_provider.extract_keywords(query)
+        #
+        # embedder = get_specter2_query_embedder()
+        # embeddings = embedder.embed_batch(keywords)
+        #
+        # # Additionally embed the original query -> could also be used for searching
+        # embeddings.append(embedder.embed_one(query))
+        #
+        # TODO: Use `embeddings` to perform a vector similarity search against
+        # the `paper.embedding` column (pgvector) and return real results.
+
+        # Simple placeholder: return the 5 most recently fetched papers
+        papers_db = (
+            db.query(PaperModel)
+            .order_by(desc(PaperModel.fetched_at))
+            .limit(10)  # fetch a few more and filter below
+            .all()
         )
 
-        return SearchResponse(papers=[example_paper])
+        # Manually map ORM objects to schema to handle slight type/field
+        # differences (e.g. authors JSON shape, missing URLs, nullable DOIs).
+        paper_schemas: list[PaperSchema] = []
+        for p in papers_db:
+            # Only require a title; DOI and URLs may be missing or unreliable.
+            if not p.title:
+                continue
+
+            authors_value = None
+            if isinstance(p.authors, dict):
+                authors_value = p.authors
+            elif isinstance(p.authors, list):
+                # Example format: [["Last", "First", "Middle"], ...]
+                formatted = []
+                for item in p.authors:
+                    if isinstance(item, (list, tuple)) and item:
+                        last = item[0] or ""
+                        first = item[1] if len(item) > 1 else ""
+                        middle = item[2] if len(item) > 2 else ""
+                        name = " ".join(part for part in [first, middle, last] if part)
+                        if name:
+                            formatted.append(name)
+                if formatted:
+                    authors_value = {str(idx): name for idx, name in enumerate(formatted)}
+
+            paper_schemas.append(
+                PaperSchema(
+                    paper_id=p.paper_id,
+                    doi=p.doi,
+                    source=str(getattr(p.source, "value", p.source)),
+                    paper_type=str(getattr(p.paper_type, "value", p.paper_type)),
+                    title=p.title,
+                    authors=authors_value,
+                    abstract=p.abstract,
+                    published_at=p.published_at,
+                    pdf_url=None,
+                    url=None,
+                    fetched_at=p.fetched_at,
+                )
+            )
+
+            if len(paper_schemas) >= 5:
+                break
+
+        return SearchResponse(papers=paper_schemas)