Merge branch 'temporary_branch' of https://github.com/QuickfoxConsulting/privateGPT into temporary_branch

.env

@@ -5,7 +5,7 @@ DB_HOST=localhost
 DB_USER=postgres
 DB_PORT=5432
 DB_PASSWORD=admin
-DB_NAME=openai
+DB_NAME=QuickGpt
 
 SUPER_ADMIN_EMAIL=superadmin@email.com
 SUPER_ADMIN_PASSWORD=supersecretpassword

Dockerfile.external

@@ -33,7 +33,7 @@ COPY --chown=worker private_gpt/ private_gpt
 COPY --chown=worker fern/ fern
 COPY --chown=worker *.yaml *.md ./
 COPY --chown=worker scripts/ scripts
-RUN poetry run python scripts/setup
+RUN 
 
 ENV PYTHONPATH="$PYTHONPATH:/private_gpt/"
 

alembic/versions/57167fd38273_changes.py

@@ -0,0 +1,53 @@
+"""Changes
+
+Revision ID: 57167fd38273
+Revises: 
+Create Date: 2024-04-07 12:33:40.975147
+
+"""
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+
+
+# revision identifiers, used by Alembic.
+revision: str = '57167fd38273'
+down_revision: Union[str, None] = None
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.create_table('chat_history',
+    sa.Column('conversation_id', sa.UUID(), nullable=False),
+    sa.Column('title', sa.String(length=255), nullable=True),
+    sa.Column('created_at', sa.DateTime(), nullable=True),
+    sa.Column('updated_at', sa.DateTime(), nullable=True),
+    sa.Column('user_id', sa.Integer(), nullable=True),
+    sa.Column('_title_generated', sa.Boolean(), nullable=True),
+    sa.ForeignKeyConstraint(['user_id'], ['users.id'], ),
+    sa.PrimaryKeyConstraint('conversation_id')
+    )
+    op.create_table('chat_items',
+    sa.Column('id', sa.Integer(), nullable=False),
+    sa.Column('sender', sa.String(length=225), nullable=False),
+    sa.Column('content', sa.JSON(), nullable=True),
+    sa.Column('created_at', sa.DateTime(), nullable=True),
+    sa.Column('updated_at', sa.DateTime(), nullable=True),
+    sa.Column('like', sa.Boolean(), nullable=True),
+    sa.Column('conversation_id', sa.UUID(), nullable=False),
+    sa.ForeignKeyConstraint(['conversation_id'], ['chat_history.conversation_id'], ),
+    sa.PrimaryKeyConstraint('id')
+    )
+    # op.create_unique_constraint('unique_user_role', 'user_roles', ['user_id', 'role_id', 'company_id'])
+    # ### end Alembic commands ###
+
+
+def downgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    # op.drop_constraint('unique_user_role', 'user_roles', type_='unique')
+    op.drop_table('chat_items')
+    op.drop_table('chat_history')
+    # ### end Alembic commands ###

alembic/versions/5f1c1d3934a1_changes_content_to_json.py

@@ -0,0 +1,39 @@
+"""Changes content to json
+
+Revision ID: 5f1c1d3934a1
+Revises: eb18396f592a
+Create Date: 2024-04-07 12:29:18.009919
+
+"""
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+
+
+# revision identifiers, used by Alembic.
+revision: str = '5f1c1d3934a1'
+down_revision: Union[str, None] = 'eb18396f592a'
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.alter_column('chat_items', 'content',
+               existing_type=sa.TEXT(),
+               type_=sa.JSON(),
+               existing_nullable=True,
+               )
+    # op.create_unique_constraint('unique_user_role', 'user_roles', ['user_id', 'role_id', 'company_id'])
+    # ### end Alembic commands ###
+
+
+def downgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    # op.drop_constraint('unique_user_role', 'user_roles', type_='unique')
+    op.alter_column('chat_items', 'content',
+               existing_type=sa.JSON(),
+               type_=sa.TEXT(),
+               existing_nullable=True)
+    # ### end Alembic commands ###

alembic/versions/9957402017dc_create_chat_history_and_items.py

@@ -0,0 +1,52 @@
+"""create chat history and items
+
+Revision ID: 9957402017dc
+Revises: 
+Create Date: 2024-04-04 11:31:53.261330
+
+"""
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+
+
+# revision identifiers, used by Alembic.
+revision: str = '9957402017dc'
+down_revision: Union[str, None] = None
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    op.create_table('chat_history',
+    sa.Column('conversation_id', sa.UUID(), nullable=False),
+    sa.Column('title', sa.String(length=255), nullable=True),
+    sa.Column('created_at', sa.DateTime(), nullable=True),
+    sa.Column('updated_at', sa.DateTime(), nullable=True),
+    sa.Column('user_id', sa.Integer(), nullable=True),
+    sa.ForeignKeyConstraint(['user_id'], ['users.id'], ),
+    sa.PrimaryKeyConstraint('conversation_id')
+    )
+    op.create_table('chat_items',
+    sa.Column('id', sa.Integer(), nullable=False),
+    sa.Column('sender', sa.String(length=225), nullable=False),
+    sa.Column('content', sa.Text(), nullable=True),
+    sa.Column('created_at', sa.DateTime(), nullable=True),
+    sa.Column('updated_at', sa.DateTime(), nullable=True),
+    sa.Column('like', sa.Boolean(), nullable=True),
+    sa.Column('conversation_id', sa.UUID(), nullable=False),
+    sa.ForeignKeyConstraint(['conversation_id'], ['chat_history.conversation_id'], ),
+    sa.PrimaryKeyConstraint('id')
+    )
+    # op.create_unique_constraint('unique_user_role', 'user_roles', ['user_id', 'role_id', 'company_id'])
+    # ### end Alembic commands ###
+
+
+def downgrade() -> None:
+    # ### commands auto generated by Alembic - please adjust! ###
+    # op.drop_constraint('unique_user_role', 'user_roles', type_='unique')
+    op.drop_table('chat_items')
+    op.drop_table('chat_history')
+    # ### end Alembic commands ###

alembic/versions/b7b896502e8e_update.py

@@ -1,38 +0,0 @@
-"""update
-
-Revision ID: b7b896502e8e
-Revises: 
-Create Date: 2024-03-17 15:07:10.795935
-
-"""
-from typing import Sequence, Union
-
-from alembic import op
-import sqlalchemy as sa
-
-
-# revision identifiers, used by Alembic.
-revision: str = 'b7b896502e8e'
-down_revision: Union[str, None] = None
-branch_labels: Union[str, Sequence[str], None] = None
-depends_on: Union[str, Sequence[str], None] = None
-
-
-def upgrade() -> None:
-    # ### commands auto generated by Alembic - please adjust! ###
-    op.drop_constraint('document_department_association_department_id_fkey', 'document_department_association', type_='foreignkey')
-    op.drop_constraint('document_department_association_document_id_fkey', 'document_department_association', type_='foreignkey')
-    op.create_foreign_key(None, 'document_department_association', 'document', ['document_id'], ['id'], onupdate='CASCADE', ondelete='CASCADE')
-    op.create_foreign_key(None, 'document_department_association', 'departments', ['department_id'], ['id'], onupdate='CASCADE', ondelete='CASCADE')
-    # op.create_unique_constraint('unique_user_role', 'user_roles', ['user_id', 'role_id', 'company_id'])
-    # ### end Alembic commands ###
-
-
-def downgrade() -> None:
-    # ### commands auto generated by Alembic - please adjust! ###
-    # op.drop_constraint('unique_user_role', 'user_roles', type_='unique')
-    op.drop_constraint(None, 'document_department_association', type_='foreignkey')
-    op.drop_constraint(None, 'document_department_association', type_='foreignkey')
-    op.create_foreign_key('document_department_association_document_id_fkey', 'document_department_association', 'document', ['document_id'], ['id'])
-    op.create_foreign_key('document_department_association_department_id_fkey', 'document_department_association', 'departments', ['department_id'], ['id'])
-    # ### end Alembic commands ###

alembic/versions/eb18396f592a_create_title_event.py

@@ -1,8 +1,8 @@
-"""Update is_enabled to false by default
+"""create title event
 
-Revision ID: 14281ff34686
-Revises: b7b896502e8e
-Create Date: 2024-03-18 16:33:43.133458
+Revision ID: eb18396f592a
+Revises: 9957402017dc
+Create Date: 2024-04-04 11:47:55.600187
 
 """
 from typing import Sequence, Union
@@ -12,21 +12,21 @@ import sqlalchemy as sa
 
 
 # revision identifiers, used by Alembic.
-revision: str = '14281ff34686'
-down_revision: Union[str, None] = 'b7b896502e8e'
+revision: str = 'eb18396f592a'
+down_revision: Union[str, None] = '9957402017dc'
 branch_labels: Union[str, Sequence[str], None] = None
 depends_on: Union[str, Sequence[str], None] = None
 
 
 def upgrade() -> None:
     # ### commands auto generated by Alembic - please adjust! ###
+    op.add_column('chat_history', sa.Column('_title_generated', sa.Boolean(), nullable=True))
     # op.create_unique_constraint('unique_user_role', 'user_roles', ['user_id', 'role_id', 'company_id'])
     # ### end Alembic commands ###
-    pass
 
 
 def downgrade() -> None:
     # ### commands auto generated by Alembic - please adjust! ###
     # op.drop_constraint('unique_user_role', 'user_roles', type_='unique')
-    pass
+    op.drop_column('chat_history', '_title_generated')
     # ### end Alembic commands ###

alembic_readme.md

@@ -4,4 +4,8 @@
 
 `alembic revision --autogenerate -m "Create user model"` # first migration
 
-`alembic upgrade 66b63a` # reflect migration on database (here 66b63a) is ssh value
+`alembic upgrade 66b63a` # reflect migration on database (here 66b63a) is ssh value
+
+
+## Local installation
+`poetry install --extras "ui llms-llama-cpp embeddings-huggingface vector-stores-qdrant rerank-sentence-transformers"`

poetry.lock

@@ -1,4 +1,4 @@
-# This file is automatically @generated by Poetry 1.8.2 and should not be changed by hand.
+# This file is automatically @generated by Poetry 1.7.1 and should not be changed by hand.
 
 [[package]]
 name = "aiofiles"
@@ -6511,6 +6511,16 @@ h2 = ["h2 (>=4,<5)"]
 socks = ["pysocks (>=1.5.6,!=1.5.7,<2.0)"]
 zstd = ["zstandard (>=0.18.0)"]
 
+[[package]]
+name = "uuid"
+version = "1.30"
+description = "UUID object and generation functions (Python 2.3 or higher)"
+optional = false
+python-versions = "*"
+files = [
+    {file = "uuid-1.30.tar.gz", hash = "sha256:1f87cc004ac5120466f36c5beae48b4c48cc411968eed0eaecd3da82aa96193f"},
+]
+
 [[package]]
 name = "uvicorn"
 version = "0.28.0"
@@ -7147,4 +7157,4 @@ vector-stores-qdrant = ["llama-index-vector-stores-qdrant"]
 [metadata]
 lock-version = "2.0"
 python-versions = ">=3.11,<3.12"
-content-hash = "b0715db49b2b01f8b6c1b3d122d93de58ef052be999114390fab9fa19cc9e794"
+content-hash = "57d64e26ba5f7100c0f6d61135808b190a9f6b8c95a6407b984b7cbea244d5f5"

private_gpt/components/embedding/embedding_component.py

@@ -3,7 +3,7 @@ import logging
 from injector import inject, singleton
 from llama_index.core.embeddings import BaseEmbedding, MockEmbedding
 
-from private_gpt.paths import models_cache_path
+from private_gpt.paths import models_cache_path, models_path
 from private_gpt.settings.settings import Settings
 
 logger = logging.getLogger(__name__)
@@ -27,7 +27,6 @@ class EmbeddingComponent:
                     raise ImportError(
                         "Local dependencies not found, install with `poetry install --extras embeddings-huggingface`"
                     ) from e
-
                 self.embedding_model = HuggingFaceEmbedding(
                     model_name=settings.huggingface.embedding_hf_model_name,
                     cache_folder=str(models_cache_path),

private_gpt/components/llm/llm_component.py

@@ -11,8 +11,7 @@ from private_gpt.paths import models_cache_path, models_path
 from private_gpt.settings.settings import Settings
 
 logger = logging.getLogger(__name__)
-
-
+local_path = models_path / "tokenizer/Mistral-7B-Instruct-v0.1"
 @singleton
 class LLMComponent:
     llm: LLM
@@ -23,8 +22,7 @@ class LLMComponent:
         if settings.llm.tokenizer:
             set_global_tokenizer(
                 AutoTokenizer.from_pretrained(
-                    pretrained_model_name_or_path=settings.llm.tokenizer,
-                    cache_dir=str(models_cache_path),
+                    local_path
                 )
             )
 

private_gpt/launcher.py

@@ -54,6 +54,7 @@ def create_app(root_injector: Injector) -> FastAPI:
                            "http://192.168.1.98", "http://192.168.1.98:5173", "http://localhost:3000","https://globaldocquery.gibl.com.np/", "http://127.0.0.1/", "http://localhost/", 
                            "http://localhost:80", "http://192.168.1.131", 'http://192.168.1.131:3000'
                            , "http://192.168.1.127", 'http://192.168.1.127:3000'
+                           , "http://192.168.1.70", 'http://192.168.1.70:3000'
                            ],
             allow_methods=["DELETE", "GET", "POST", "PUT", "OPTIONS", "PATCH"],
             allow_headers=["*"],

private_gpt/server/completions/completions_router.py

@@ -4,10 +4,11 @@ from llama_index.llms import ChatMessage, ChatResponse, MessageRole
 from fastapi import APIRouter, Depends, Request, Security, HTTPException, status
 from private_gpt.server.ingest.ingest_service import IngestService
 from pydantic import BaseModel
+from typing import List, Dict, Any
 from sqlalchemy.orm import Session
 import traceback
 import logging
-
+import json
 logger = logging.getLogger(__name__)
 
 from starlette.responses import StreamingResponse
@@ -20,13 +21,13 @@ from private_gpt.open_ai.openai_models import (
 from private_gpt.server.chat.chat_router import ChatBody, chat_completion
 from private_gpt.server.utils.auth import authenticated
 from private_gpt.users.api import deps
-from pydantic import Optional
-
+from private_gpt.users import crud, models, schemas
+import uuid
 completions_router = APIRouter(prefix="/v1", dependencies=[Depends(authenticated)])
 
 
 class CompletionsBody(BaseModel):
-    conversation_id: Optional[int]
+    conversation_id: uuid.UUID
     prompt: str
     system_prompt: str | None = None
     use_context: bool = False
@@ -38,6 +39,7 @@ class CompletionsBody(BaseModel):
         "json_schema_extra": {
             "examples": [
                 {
+                    "conversation_id": 123,
                     "prompt": "How do you fry an egg?",
                     "system_prompt": "You are a rapper. Always answer with a rap.",
                     "stream": False,
@@ -49,6 +51,9 @@ class CompletionsBody(BaseModel):
     }
 
 
+class ChatContentCreate(BaseModel):
+    content: Dict[str, Any]
+
 # @completions_router.post(
 #     "/completions",
 #     response_model=None,
@@ -97,6 +102,14 @@ class CompletionsBody(BaseModel):
 #     )
 #     return chat_completion(request, chat_body)
 
+def create_chat_item(db, sender, content, conversation_id):
+    chat_item_create = schemas.ChatItemCreate(
+            sender=sender,
+            content=content,
+            conversation_id=conversation_id
+        )
+    return crud.chat_item.create(db, obj_in=chat_item_create)
+
 
 @completions_router.post(
     "/chat",
@@ -135,97 +148,59 @@ async def prompt_completion(
             raise HTTPException(status_code=status.HTTP_404_NOT_FOUND,
                                 detail=f"No documents uploaded for your department.")
         docs_list = [document.filename for document in documents]
-        print("DOCUMENTS ASSIGNED TO THIS DEPARTMENTS: ", docs_list)
         docs_ids = []
         for filename in docs_list:
             doc_id = service.get_doc_ids_by_filename(filename)
             docs_ids.extend(doc_id)
         body.context_filter = {"docs_ids": docs_ids}
 
+        chat_history = crud.chat.get_by_id(
+            db, id=body.conversation_id
+        )
+        if (chat_history is None) and (chat_history.user_id != current_user.id):
+            raise HTTPException(
+                status_code=404, detail="Chat history not found")
+        
+        user_message = OpenAIMessage(content=body.prompt, role="user")
+        user_message = user_message.model_dump(mode="json")
+
+        user_message_json = {
+            'text': body.prompt,
+        }
+        create_chat_item(db, "user", user_message_json , body.conversation_id)
+        
+        messages = [user_message]
+
+        if body.system_prompt:
+            messages.insert(0, OpenAIMessage(
+                content=body.system_prompt, role="system"))
+
+        chat_body = ChatBody(
+            messages=messages,
+            use_context=body.use_context,
+            stream=body.stream,
+            include_sources=body.include_sources,
+            context_filter=body.context_filter,
+        )
+        log_audit(
+            model='Chat', 
+            action='Chat',
+            details={
+                "query": body.prompt,
+                'user': current_user.username,
+                }, 
+            user_id=current_user.id
+        )
+        
+        chat_response = await chat_completion(request, chat_body)
+        ai_response = chat_response.model_dump(mode="json")
+        create_chat_item(db, "assistant", ai_response, body.conversation_id)
+        return chat_response
+    
     except Exception as e:
         print(traceback.format_exc())
         logger.error(f"There was an error: {str(e)}")
         raise HTTPException(
             status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
             detail="Internal Server Error",
-        )
-
-    if body.conversation_id:
-        chat_history = crud.chat.get_by_id(db, id=body.conversation_id)
-        if chat_history is None or chat_history.user_id != current_user.id:
-            raise HTTPException(
-                status_code=404, detail="Chat history not found")
-    else:
-        chat_create_in = schemas.ChatCreate(user_id=current_user.id)
-        chat_history = crud.chat.create(db=db, obj_in=chat_create_in)
-
-    _history = chat_history.messages or []
-
-    def build_history() -> list[ChatMessage]:
-        history_messages: list[ChatMessage] = []
-        for interaction in _history:
-            user_message = interaction.get("user", "")
-            ai_message = interaction.get("ai", "")
-            if user_message:
-                history_messages.append(
-                    ChatMessage(
-                        content=user_message,
-                        role=MessageRole.USER
-                    )
-                )
-            if ai_message:
-                history_messages.append(
-                    ChatMessage(
-                        content=ai_message,
-                        role=MessageRole.ASSISTANT
-                    )
-                )
-
-        # max 20 messages to try to avoid context overflow
-        return history_messages[:20]
-
-    # Prepare new messages
-    new_messages = []
-
-    if body.prompt:
-        new_messages.append(OpenAIMessage(content=body.prompt, role="user"))
-    if body.system_prompt:
-        new_messages.insert(0, OpenAIMessage(
-            content=body.system_prompt, role="system"))
-
-    # Update chat history with new user messages
-    if new_messages:
-        new_message = ChatMessage(content=new_messages, role=MessageRole.USER)
-        _history.append(new_message.dict())
-
-    # Process chat completion
-    chat_body = ChatBody(
-        messages=build_history(),
-        use_context=body.use_context,
-        stream=body.stream,
-        include_sources=body.include_sources,
-        context_filter=body.context_filter,
-    )
-
-    ai_response = await chat_completion(request, chat_body)
-
-    # Update chat history with AI response
-    if ai_response.messages:
-        ai_message = OpenAIMessage(
-            content=ai_response.messages, role="assistant")
-        _history.append(ai_message.dict())
-
-    # Update chat history in the database
-    chat_obj_in = schemas.ChatUpdate(messages=build_history())
-    crud.chat.update_messages(db, db_obj=chat_history, obj_in=chat_obj_in)
-
-    return ai_response
-    # log_audit(
-    #     model='Chat', 
-    #     action='Chat',
-    #     details={
-    #         "query": body.prompt,
-    #         'user': current_user.username,
-    #         }, 
-    #     user_id=current_user.id
-    # )
+        )

private_gpt/users/api/v1/api.py

@@ -1,5 +1,5 @@
 from private_gpt.users.api import deps
-from private_gpt.users.api.v1.routers import auth, roles, user_roles, users, subscriptions, companies, departments, documents, audits
+from private_gpt.users.api.v1.routers import auth, roles, user_roles, users, subscriptions, companies, departments, documents, audits, chat_history
 from fastapi import APIRouter
 
 api_router = APIRouter(prefix="/v1")
@@ -13,4 +13,5 @@ api_router.include_router(subscriptions.router)
 api_router.include_router(departments.router)
 api_router.include_router(documents.router)
 api_router.include_router(audits.router)
+api_router.include_router(chat_history.router)
 

private_gpt/users/api/v1/routers/chat_history.py

@@ -0,0 +1,125 @@
+import logging
+import traceback
+import uuid
+from sqlalchemy.orm import Session
+from fastapi.responses import JSONResponse
+from fastapi import APIRouter, Depends, HTTPException, status, Security
+
+from private_gpt.users.api import deps
+from private_gpt.users import crud, models, schemas
+
+logger = logging.getLogger(__name__)
+router = APIRouter(prefix="/c", tags=["Chat Histories"])
+
+
+@router.get("", response_model=list[schemas.ChatHistory])
+def list_chat_histories(
+    db: Session = Depends(deps.get_db),
+    skip: int = 0,
+    limit: int = 100,
+    current_user: models.User = Security(
+        deps.get_current_user,
+    ),
+) -> list[schemas.ChatHistory]:
+    """
+    Retrieve a list of chat histories with pagination support.
+    """
+    try:
+        chat_histories = crud.chat.get_chat_history(
+            db, skip=skip, limit=limit)
+        return chat_histories
+    except Exception as e:
+        print(traceback.format_exc())
+        logger.error(f"Error listing chat histories: {str(e)}")
+        raise HTTPException(
+            status_code=500,
+            detail="Internal Server Error",
+        )
+
+
+@router.post("/create", response_model=schemas.ChatHistory)
+def create_chat_history(
+    db: Session = Depends(deps.get_db),
+    current_user: models.User = Security(
+        deps.get_current_user,
+    ),
+) -> schemas.ChatHistory:
+    """
+    Create a new chat history
+    """
+    try:
+        chat_history_in = schemas.CreateChatHistory(
+            user_id= current_user.id
+        )
+        chat_history = crud.chat.create(
+            db=db, obj_in=chat_history_in)
+        return chat_history
+    except Exception as e:
+        print(traceback.format_exc())
+        logger.error(f"Error creating chat history: {str(e)}")
+        raise HTTPException(
+            status_code=500,
+            detail="Internal Server Error",
+        )
+
+
+@router.get("/{conversation_id}", response_model=schemas.ChatHistory)
+def read_chat_history(
+    conversation_id: uuid.UUID,
+    db: Session = Depends(deps.get_db),
+    current_user: models.User = Security(
+        deps.get_current_user,
+    ),
+) -> schemas.ChatHistory:
+    """
+    Read a chat history by ID
+    """
+    try:
+        chat_history = crud.chat.get_by_id(db, id=conversation_id)
+        if chat_history is None or chat_history.user_id != current_user.id:
+            raise HTTPException(
+                status_code=404, detail="Chat history not found")
+        return chat_history
+    except Exception as e:
+        print(traceback.format_exc())
+        logger.error(f"Error reading chat history: {str(e)}")
+        raise HTTPException(
+            status_code=500,
+            detail="Internal Server Error",
+        )
+
+
+@router.post("/delete")
+def delete_chat_history(
+    chat_history_in: schemas.ChatDelete,
+    db: Session = Depends(deps.get_db),
+    current_user: models.User = Security(
+        deps.get_current_user,
+    ),
+):
+    """
+    Delete a chat history by ID
+    """
+    try:
+        chat_history_id = chat_history_in.conversation_id
+        chat_history = crud.chat.get(db, id=chat_history_id)
+        if chat_history is None or chat_history.user_id != current_user.id:
+            raise HTTPException(
+                status_code=404, detail="Chat history not found")
+
+        crud.chat.remove(db=db, id=chat_history_id)
+        return JSONResponse(
+            status_code=status.HTTP_200_OK,
+            content={
+                "message": "Chat history deleted successfully",
+            },
+        )
+    except Exception as e:
+        print(traceback.format_exc())
+        logger.error(f"Error deleting chat history: {str(e)}")
+        raise HTTPException(
+            status_code=500,
+            detail="Internal Server Error",
+        )
+
+

private_gpt/users/crud/__init__.py

@@ -6,4 +6,4 @@ from .subscription_crud import subscription
 from .document_crud import documents
 from .department_crud import department
 from .audit_crud import audit
-from .chathistory_crud import chat
+from .chat_crud import chat, chat_item

private_gpt/users/crud/chat_crud.py

@@ -0,0 +1,32 @@
+from sqlalchemy.sql.expression import desc, asc
+from typing import Optional, List, Union, Dict, Any
+from fastapi import HTTPException, status
+from sqlalchemy.orm import Session
+from sqlalchemy.exc import IntegrityError
+import uuid
+from private_gpt.users.crud.base import CRUDBase
+from private_gpt.users.models.chat import ChatHistory, ChatItem
+from private_gpt.users.schemas.chat import ChatHistoryCreate, ChatHistoryCreate, ChatItemCreate, ChatItemUpdate
+
+
+class CRUDChat(CRUDBase[ChatHistory, ChatHistoryCreate, ChatHistoryCreate]):
+    def get_by_id(self, db: Session, *, id: uuid.UUID) -> Optional[ChatHistory]:
+        return db.query(self.model).filter(ChatHistory.conversation_id == id).first()
+
+    def get_chat_history(
+            self, db: Session, *, skip: int = 0, limit: int = 100
+        ) -> List[ChatHistory]:
+            return (
+                db.query(self.model)
+                .order_by(desc(getattr(ChatHistory, 'created_at')))
+                .offset(skip)
+                .limit(limit)
+                .all()
+            )
+
+class CRUDChatItem(CRUDBase[ChatItem, ChatItemCreate, ChatItemUpdate]):
+    pass
+
+
+chat = CRUDChat(ChatHistory)
+chat_item = CRUDChatItem(ChatItem)

private_gpt/users/crud/document_crud.py

@@ -35,9 +35,10 @@ class CRUDDocuments(CRUDBase[Document, DocumentCreate, DocumentUpdate]):
                 .join(document_department_association)
                 .join(Department)
                 .filter(document_department_association.c.department_id == department_id)
+                .order_by(desc(getattr(Document, 'uploaded_at')))
                 .offset(skip)
                 .limit(limit)
-                .all().order_by(desc(getattr(Document, 'uploaded_at')))
+                .all()
             )
     
     def get_files_to_verify(

private_gpt/users/models/__init__.py

@@ -7,4 +7,4 @@ from .subscription import Subscription
 from .department import Department
 from .audit import Audit
 from .document_department import document_department_association
-from .chat_history import ChatHistory
+from .chat import ChatHistory, ChatItem

private_gpt/users/models/chat.py

@@ -0,0 +1,71 @@
+import uuid
+from datetime import datetime
+from sqlalchemy.orm import relationship
+from sqlalchemy.dialects.postgresql import UUID 
+from sqlalchemy import Column, Integer, String, DateTime, ForeignKey, Text, Boolean, event, JSON
+
+from private_gpt.users.db.base_class import Base
+
+class ChatHistory(Base):
+    """Models a chat history table"""
+
+    __tablename__ = "chat_history"
+
+    conversation_id = Column(UUID(as_uuid=True), primary_key=True, default=uuid.uuid4)
+    title = Column(String(255), nullable=True)
+    created_at = Column(DateTime, default=datetime.now)
+    updated_at = Column(DateTime, default=datetime.now,
+                        onupdate=datetime.now)
+    user_id = Column(Integer, ForeignKey("users.id"))
+    user = relationship("User", back_populates="chat_histories")
+    chat_items = relationship(
+        "ChatItem", back_populates="chat_history", cascade="all, delete-orphan")
+    _title_generated = Column(Boolean, default=False)
+
+    def __init__(self, user_id, chat_items=None, **kwargs):
+        super().__init__(**kwargs)
+        self.user_id = user_id
+        self.chat_items = chat_items or []
+        self.generate_title()
+
+    def generate_title(self):
+        user_chat_items = [
+            item for item in self.chat_items if item.role == "user"]
+        if user_chat_items:
+            first_user_chat_item = user_chat_items[0]
+            self.title = first_user_chat_item.content[:30]
+        else:
+            self.title = str(self.conversation_id)
+
+    def __repr__(self):
+        """Returns string representation of model instance"""
+        return f"<ChatHistory {self.conversation_id!r}>"
+
+
+class ChatItem(Base):
+    """Models a chat item table"""
+
+    __tablename__ = "chat_items"
+
+    id = Column(Integer, nullable=False, primary_key=True)
+    sender = Column(String(225), nullable=False)
+    content = Column(JSON, nullable=True)
+    created_at = Column(DateTime, default=datetime.now)
+    updated_at = Column(DateTime, default=datetime.now,
+                        onupdate=datetime.now)
+    like = Column(Boolean, default=True)
+    conversation_id = Column(UUID(as_uuid=True), ForeignKey(
+        "chat_history.conversation_id"), nullable=False)
+    chat_history = relationship("ChatHistory", back_populates="chat_items")
+
+    def __repr__(self):
+        """Returns string representation of model instance"""
+        return f"<ChatItem {self.id!r}>"
+
+
+@event.listens_for(ChatHistory, "after_insert")
+def receive_after_insert(mapper, connection, target):
+    """Update title after insertion to reflect the conversation_id"""
+    if not target._title_generated:
+        target.generate_title()
+        target._title_generated = True

private_gpt/users/schemas/__init__.py

@@ -1,10 +1,25 @@
 from .role import Role, RoleCreate, RoleInDB, RoleUpdate
 from .token import TokenSchema, TokenPayload
-from .user import User, UserCreate, UserInDB, UserUpdate, UserBaseSchema, Profile, UsernameUpdate, DeleteUser, UserAdminUpdate, UserAdmin, PasswordUpdate
+from .user import (
+    User, UserCreate, UserInDB, UserUpdate, UserBaseSchema, Profile, 
+    UsernameUpdate, DeleteUser, UserAdminUpdate, UserAdmin, PasswordUpdate
+)
 from .user_role import UserRole, UserRoleCreate, UserRoleInDB, UserRoleUpdate
-from .subscription import Subscription, SubscriptionBase, SubscriptionCreate, SubscriptionUpdate
+from .subscription import (
+    Subscription, SubscriptionBase, SubscriptionCreate, SubscriptionUpdate
+)
 from .company import Company, CompanyBase, CompanyCreate, CompanyUpdate
-from .documents import Document, DocumentCreate, DocumentsBase, DocumentUpdate, DocumentList, DepartmentList, DocumentEnable, DocumentDepartmentUpdate, DocumentCheckerUpdate, DocumentMakerCreate, DocumentDepartmentList, DocumentView, DocumentVerify, DocumentFilter
-from .department import Department, DepartmentCreate, DepartmentUpdate, DepartmentAdminCreate, DepartmentDelete
+from .documents import (
+    Document, DocumentCreate, DocumentsBase, DocumentUpdate, DocumentList, 
+    DepartmentList, DocumentEnable, DocumentDepartmentUpdate, DocumentCheckerUpdate, 
+    DocumentMakerCreate, DocumentDepartmentList, DocumentView, DocumentVerify, 
+    DocumentFilter
+)
+from .department import (
+    Department, DepartmentCreate, DepartmentUpdate, DepartmentAdminCreate, DepartmentDelete
+)
 from .audit import AuditBase, AuditCreate, AuditUpdate, Audit, GetAudit
-from .chat_history import Chat, ChatBase, ChatCreate, ChatDelete, ChatUpdate, ChatMessages
+from .chat import (
+    ChatHistory, ChatHistoryBase, ChatHistoryCreate, ChatHistoryUpdate, ChatDelete,
+    ChatItem, ChatItemBase, ChatItemCreate, ChatItemUpdate, CreateChatHistory
+)

private_gpt/users/schemas/chat.py

@@ -0,0 +1,56 @@
+from datetime import datetime
+from typing import List, Optional, Union, Dict
+from pydantic import BaseModel, Json
+import uuid
+
+class ChatItemBase(BaseModel):
+    conversation_id: uuid.UUID
+    sender: str
+    content: Union[str, Dict]
+
+
+class ChatItemCreate(ChatItemBase):
+    pass
+
+class ChatItemUpdate(ChatItemBase):
+    like: Optional[bool]
+
+
+class ChatItem(ChatItemBase):
+    id: int
+    created_at: datetime
+    updated_at: datetime
+
+    class Config:
+        orm_mode = True
+
+
+class ChatHistoryBase(BaseModel):
+    user_id: int
+    title: Optional[str]
+
+
+class ChatHistoryCreate(ChatHistoryBase):
+    chat_items: Optional[List[ChatItemCreate]]
+
+class ChatHistoryUpdate(ChatHistoryBase):
+    updated_at: datetime
+    chat_items: Optional[List[ChatItemCreate]]
+
+class Chat(BaseModel):
+    conversation_id: uuid.UUID
+
+class ChatHistory(ChatHistoryBase):
+    conversation_id: uuid.UUID
+    created_at: datetime
+    updated_at: datetime
+    chat_items: List[ChatItem]
+
+    class Config:
+        orm_mode = True
+
+class ChatDelete(BaseModel):
+    conversation_id: uuid.UUID
+
+class CreateChatHistory(BaseModel):
+    user_id: int

pyproject.toml

@@ -53,6 +53,7 @@ gradio = {version ="^4.19.2", optional = true}
 aiofiles = "^23.2.1"
 timm = "^0.9.16"
 fastapi-filter = {extras = ["sqlalchemy"], version = "^1.1.0"}
+uuid = "^1.30"
 
 [tool.poetry.extras]
 ui = ["gradio"]

scripts/setup

@@ -10,7 +10,7 @@ from private_gpt.settings.settings import settings
 
 resume_download = True
 if __name__ == '__main__':
-    parser = argparse.ArgumentParser(prog='Setup: Download models from huggingface')
+    parser = argparse.ArgumentParser(prog='Setup: Download models from Hugging Face')
     parser.add_argument('--resume', default=True, action=argparse.BooleanOptionalAction, help='Enable/Disable resume_download options to restart the download progress interrupted')
     args = parser.parse_args()
     resume_download = args.resume
@@ -40,10 +40,10 @@ print("LLM model downloaded!")
 
 # Download Tokenizer
 print(f"Downloading tokenizer {settings().llm.tokenizer}")
+local_path = models_path / "tokenizer/Mistral-7B-Instruct-v0.1"
 AutoTokenizer.from_pretrained(
-    pretrained_model_name_or_path=settings().llm.tokenizer,
-    cache_dir=models_cache_path,
+    local_path
 )
 print("Tokenizer downloaded!")
 
-print("Setup done")
+print("Setup done")

settings-local.yaml

@@ -8,16 +8,24 @@ llm:
   context_window: 3900
   tokenizer: mistralai/Mistral-7B-Instruct-v0.2
 
+ 
 llamacpp:
-  prompt_style: "mistral"
-  llm_hf_repo_id: TheBloke/Mistral-7B-Instruct-v0.2-GGUF
-  llm_hf_model_file: mistral-7b-instruct-v0.2.Q4_K_M.gguf
+  prompt_style: "chatml"
+  llm_hf_repo_id: TheBloke/OpenHermes-2.5-Mistral-7B-GGUF
+  llm_hf_model_file: openhermes-2.5-mistral-7b.Q5_K_M.gguf
+  tfs_z: 1.0            # Tail free sampling is used to reduce the impact of less probable tokens from the output. A higher value (e.g., 2.0) will reduce the impact more, while a value of 1.0 disables this setting
+  top_k: 40             # Reduces the probability of generating nonsense. A higher value (e.g. 100) will give more diverse answers, while a lower value (e.g. 10) will be more conservative. (Default: 40)
+  top_p: 0.9            # Works together with top-k. A higher value (e.g., 0.95) will lead to more diverse text, while a lower value (e.g., 0.5) will generate more focused and conservative text. (Default: 0.9)
+  repeat_last_n: 64     # Sets how far back for the model to look back to prevent repetition. (Default: 64, 0 = disabled, -1 = num_ctx)
+  repeat_penalty: 1.1   # Sets how strongly to penalize repetitions. A higher value (e.g., 1.5) will penalize repetitions more strongly, while a lower value (e.g., 0.9) will be more lenient. (Default: 1.1)
 
 embedding:
+  # Should be matching the value above in most cases
   mode: huggingface
+  ingest_mode: simple
 
 huggingface:
-  embedding_hf_model_name: BAAI/bge-small-en-v1.5
+  embedding_hf_model_name: BAAI/bge-large-en-v1.5
 
 vectorstore:
   database: qdrant

settings.yaml

@@ -59,7 +59,7 @@ embedding:
   ingest_mode: simple
 
 huggingface:
-  embedding_hf_model_name: BAAI/bge-large-en-v1.5
+  embedding_hf_model_name: mixedbread-ai/mxbai-embed-large-v1
 
 vectorstore:
   database: qdrant