[Feature] Add the chat panel

surfsense_backend/alembic/versions/34_add_podcast_staleness_detection.py

@@ -0,0 +1,60 @@
+"""Add podcast staleness detection columns to chats and podcasts tables
+
+This feature allows the system to detect when a podcast is outdated compared to
+the current state of the chat it was generated from, enabling users to regenerate
+podcasts when needed.
+
+Revision ID: 34
+Revises: 33
+"""
+
+from collections.abc import Sequence
+
+from alembic import op
+
+# revision identifiers
+revision: str = "34"
+down_revision: str | None = "33"
+branch_labels: str | Sequence[str] | None = None
+depends_on: str | Sequence[str] | None = None
+
+
+def upgrade() -> None:
+    """Add columns only if they don't already exist (safe for re-runs)."""
+
+    # Add 'state_version' column to chats table (default 1)
+    op.execute("""
+        ALTER TABLE chats
+        ADD COLUMN IF NOT EXISTS state_version BIGINT DEFAULT 1 NOT NULL
+    """)
+
+    # Add 'chat_state_version' column to podcasts table
+    op.execute("""
+        ALTER TABLE podcasts
+        ADD COLUMN IF NOT EXISTS chat_state_version BIGINT
+    """)
+
+    # Add 'chat_id' column to podcasts table
+    op.execute("""
+        ALTER TABLE podcasts
+        ADD COLUMN IF NOT EXISTS chat_id INTEGER
+    """)
+
+
+def downgrade() -> None:
+    """Remove columns only if they exist."""
+
+    op.execute("""
+        ALTER TABLE podcasts
+        DROP COLUMN IF EXISTS chat_state_version
+    """)
+
+    op.execute("""
+        ALTER TABLE podcasts
+        DROP COLUMN IF EXISTS chat_id
+    """)
+
+    op.execute("""
+        ALTER TABLE chats
+        DROP COLUMN IF EXISTS state_version
+    """)

surfsense_backend/app/agents/podcaster/configuration.py

@@ -18,6 +18,7 @@ class Configuration:
     podcast_title: str
     user_id: str
     search_space_id: int
+    user_prompt: str | None = None
 
     @classmethod
     def from_runnable_config(

surfsense_backend/app/agents/podcaster/nodes.py

@@ -29,6 +29,7 @@ async def create_podcast_transcript(
     configuration = Configuration.from_runnable_config(config)
     user_id = configuration.user_id
     search_space_id = configuration.search_space_id
+    user_prompt = configuration.user_prompt
 
     # Get user's long context LLM
     llm = await get_user_long_context_llm(state.db_session, user_id, search_space_id)
@@ -38,7 +39,7 @@ async def create_podcast_transcript(
         raise RuntimeError(error_message)
 
     # Get the prompt
-    prompt = get_podcast_generation_prompt()
+    prompt = get_podcast_generation_prompt(user_prompt)
 
     # Create the messages
     messages = [

surfsense_backend/app/agents/podcaster/prompts.py

@@ -1,12 +1,23 @@
 import datetime
 
 
-def get_podcast_generation_prompt():
+def get_podcast_generation_prompt(user_prompt: str | None = None):
     return f"""
 Today's date: {datetime.datetime.now().strftime("%Y-%m-%d")}
 <podcast_generation_system>
 You are a master podcast scriptwriter, adept at transforming diverse input content into a lively, engaging, and natural-sounding conversation between two distinct podcast hosts. Your primary objective is to craft authentic, flowing dialogue that captures the spontaneity and chemistry of a real podcast discussion, completely avoiding any hint of robotic scripting or stiff formality. Think dynamic interplay, not just information delivery.
 
+{
+        f'''
+You **MUST** strictly adhere to the following user instruction while generating the podcast script:
+<user_instruction>
+{user_prompt}
+</user_instruction>
+'''
+        if user_prompt
+        else ""
+    }
+
 <input>
 - '<source_content>': A block of text containing the information to be discussed in the podcast. This could be research findings, an article summary, a detailed outline, user chat history related to the topic, or any other relevant raw information. The content might be unstructured but serves as the factual basis for the podcast dialogue.
 </input>

surfsense_backend/app/db.py

@@ -9,6 +9,7 @@
     ARRAY,
     JSON,
     TIMESTAMP,
+    BigInteger,
     Boolean,
     Column,
     Enum as SQLAlchemyEnum,
@@ -157,6 +158,7 @@ class Chat(BaseModel, TimestampMixin):
     title = Column(String, nullable=False, index=True)
     initial_connectors = Column(ARRAY(String), nullable=True)
     messages = Column(JSON, nullable=False)
+    state_version = Column(BigInteger, nullable=False, default=1)
 
     search_space_id = Column(
         Integer, ForeignKey("searchspaces.id", ondelete="CASCADE"), nullable=False
@@ -203,6 +205,10 @@ class Podcast(BaseModel, TimestampMixin):
     title = Column(String, nullable=False, index=True)
     podcast_transcript = Column(JSON, nullable=False, default={})
     file_location = Column(String(500), nullable=False, default="")
+    chat_id = Column(
+        Integer, ForeignKey("chats.id", ondelete="CASCADE"), nullable=True
+    )  # If generated from a chat, this will be the chat id, else null ( can be from a document or a chat )
+    chat_state_version = Column(BigInteger, nullable=True)
 
     search_space_id = Column(
         Integer, ForeignKey("searchspaces.id", ondelete="CASCADE"), nullable=False

surfsense_backend/app/routes/chats_routes.py

@@ -196,6 +196,7 @@ async def read_chats(
                 Chat.initial_connectors,
                 Chat.search_space_id,
                 Chat.created_at,
+                Chat.state_version,
             )
             .join(SearchSpace)
             .filter(SearchSpace.user_id == user.id)
@@ -258,7 +259,10 @@ async def update_chat(
         db_chat = await read_chat(chat_id, session, user)
         update_data = chat_update.model_dump(exclude_unset=True)
         for key, value in update_data.items():
+            if key == "messages":
+                db_chat.state_version = len(update_data["messages"])
             setattr(db_chat, key, value)
+
         await session.commit()
         await session.refresh(db_chat)
         return db_chat

surfsense_backend/app/routes/podcasts_routes.py

@@ -155,15 +155,19 @@ async def delete_podcast(
 
 
 async def generate_chat_podcast_with_new_session(
-    chat_id: int, search_space_id: int, podcast_title: str, user_id: int
+    chat_id: int,
+    search_space_id: int,
+    user_id: int,
+    podcast_title: str | None = None,
+    user_prompt: str | None = None,
 ):
     """Create a new session and process chat podcast generation."""
     from app.db import async_session_maker
 
     async with async_session_maker() as session:
         try:
             await generate_chat_podcast(
-                session, chat_id, search_space_id, podcast_title, user_id
+                session, chat_id, search_space_id, user_id, podcast_title, user_prompt
             )
         except Exception as e:
             import logging
@@ -211,7 +215,11 @@ async def generate_podcast(
             # Add Celery tasks for each chat ID
             for chat_id in valid_chat_ids:
                 generate_chat_podcast_task.delay(
-                    chat_id, request.search_space_id, request.podcast_title, user.id
+                    chat_id,
+                    request.search_space_id,
+                    user.id,
+                    request.podcast_title,
+                    request.user_prompt,
                 )
 
         return {
@@ -287,3 +295,27 @@ def iterfile():
         raise HTTPException(
             status_code=500, detail=f"Error streaming podcast: {e!s}"
         ) from e
+
+
+@router.get("/podcasts/by-chat/{chat_id}", response_model=PodcastRead | None)
+async def get_podcast_by_chat_id(
+    chat_id: int,
+    session: AsyncSession = Depends(get_async_session),
+    user: User = Depends(current_active_user),
+):
+    try:
+        # Get the podcast and check if user has access
+        result = await session.execute(
+            select(Podcast)
+            .join(SearchSpace)
+            .filter(Podcast.chat_id == chat_id, SearchSpace.user_id == user.id)
+        )
+        podcast = result.scalars().first()
+
+        return podcast
+    except HTTPException as he:
+        raise he
+    except Exception as e:
+        raise HTTPException(
+            status_code=500, detail=f"Error fetching podcast: {e!s}"
+        ) from e

surfsense_backend/app/schemas/chats.py

@@ -13,12 +13,14 @@ class ChatBase(BaseModel):
     initial_connectors: list[str] | None = None
     messages: list[Any]
     search_space_id: int
+    state_version: int = 1
 
 
 class ChatBaseWithoutMessages(BaseModel):
     type: ChatType
     title: str
     search_space_id: int
+    state_version: int = 1
 
 
 class ClientAttachment(BaseModel):

surfsense_backend/app/schemas/podcasts.py

@@ -10,6 +10,7 @@ class PodcastBase(BaseModel):
     podcast_transcript: list[Any]
     file_location: str = ""
     search_space_id: int
+    chat_state_version: int | None = None
 
 
 class PodcastCreate(PodcastBase):
@@ -28,4 +29,5 @@ class PodcastGenerateRequest(BaseModel):
     type: Literal["DOCUMENT", "CHAT"]
     ids: list[int]
     search_space_id: int
-    podcast_title: str = "SurfSense Podcast"
+    podcast_title: str | None = None
+    user_prompt: str | None = None

surfsense_backend/app/tasks/celery_tasks/podcast_tasks.py

@@ -38,23 +38,31 @@ def get_celery_session_maker():
 
 @celery_app.task(name="generate_chat_podcast", bind=True)
 def generate_chat_podcast_task(
-    self, chat_id: int, search_space_id: int, podcast_title: str, user_id: int
+    self,
+    chat_id: int,
+    search_space_id: int,
+    user_id: int,
+    podcast_title: str | None = None,
+    user_prompt: str | None = None,
 ):
     """
     Celery task to generate podcast from chat.
 
     Args:
         chat_id: ID of the chat to generate podcast from
         search_space_id: ID of the search space
+        user_id: ID of the user,
         podcast_title: Title for the podcast
-        user_id: ID of the user
+        user_prompt: Optional prompt from the user to guide the podcast generation
     """
     loop = asyncio.new_event_loop()
     asyncio.set_event_loop(loop)
 
     try:
         loop.run_until_complete(
-            _generate_chat_podcast(chat_id, search_space_id, podcast_title, user_id)
+            _generate_chat_podcast(
+                chat_id, search_space_id, user_id, podcast_title, user_prompt
+            )
         )
         loop.run_until_complete(loop.shutdown_asyncgens())
     finally:
@@ -63,13 +71,17 @@ def generate_chat_podcast_task(
 
 
 async def _generate_chat_podcast(
-    chat_id: int, search_space_id: int, podcast_title: str, user_id: int
+    chat_id: int,
+    search_space_id: int,
+    user_id: int,
+    podcast_title: str | None = None,
+    user_prompt: str | None = None,
 ):
     """Generate chat podcast with new session."""
     async with get_celery_session_maker()() as session:
         try:
             await generate_chat_podcast(
-                session, chat_id, search_space_id, podcast_title, user_id
+                session, chat_id, search_space_id, user_id, podcast_title, user_prompt
             )
         except Exception as e:
             logger.error(f"Error generating podcast from chat: {e!s}")