Added chat history and chat item

.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/14281ff34686_update_is_enabled_to_false_by_default.py

@@ -1,32 +0,0 @@
-"""Update is_enabled to false by default
-
-Revision ID: 14281ff34686
-Revises: b7b896502e8e
-Create Date: 2024-03-18 16:33:43.133458
-
-"""
-from typing import Sequence, Union
-
-from alembic import op
-import sqlalchemy as sa
-
-
-# revision identifiers, used by Alembic.
-revision: str = '14281ff34686'
-down_revision: Union[str, None] = 'b7b896502e8e'
-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_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
-    # ### end Alembic commands ###

alembic/versions/7bd9152cf172_create_chat_history_and_item.py

@@ -0,0 +1,52 @@
+"""Create chat history and item
+
+Revision ID: 7bd9152cf172
+Revises: 
+Create Date: 2024-04-03 16:23:24.813222
+
+"""
+from typing import Sequence, Union
+
+from alembic import op
+import sqlalchemy as sa
+
+
+# revision identifiers, used by Alembic.
+revision: str = '7bd9152cf172'
+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.Integer(), 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.Integer(), 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 ###

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/server/chat/chat_service.py

@@ -129,9 +129,7 @@ class ChatService:
             else None
         )
         system_prompt = (
-            chat_engine_input.system_message.content
-            if chat_engine_input.system_message
-            else None
+            "You can only answer questions about the provided context. If you know the answer but it is not based in the provided context, don't provide the answer, just state the answer is not in the context provided."
         )
         chat_history = (
             chat_engine_input.chat_history if chat_engine_input.chat_history else None
@@ -165,9 +163,7 @@ class ChatService:
             else None
         )
         system_prompt = (
-            chat_engine_input.system_message.content
-            if chat_engine_input.system_message
-            else None
+            "You can only answer questions about the provided context. If you know the answer but it is not based in the provided context, don't provide the answer, just state the answer is not in the context provided."
         )
         chat_history = (
             chat_engine_input.chat_history if chat_engine_input.chat_history else None

private_gpt/server/completions/completions_router.py

@@ -22,6 +22,7 @@ completions_router = APIRouter(prefix="/v1", dependencies=[Depends(authenticated
 
 
 class CompletionsBody(BaseModel):
+    conversation_id: int
     prompt: str
     system_prompt: str | None = None
     use_context: bool = False
@@ -33,6 +34,7 @@ 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,
@@ -92,6 +94,14 @@ 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",
@@ -137,33 +147,47 @@ async def prompt_completion(
             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")
+        
+        messages = [OpenAIMessage(content=body.prompt, role="user")]
+        create_chat_item(db, "user", body.prompt, body.conversation_id)
+
+        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)
+        print(chat_response)
+        create_chat_item(db, "assistant", chat_response.choices[0].message.content, 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",
-        )
-
-    messages = [OpenAIMessage(content=body.prompt, role="user")]
-    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
-    )
-    return await chat_completion(request, chat_body)
+        )

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 +1,123 @@
+import logging
+import traceback
+
+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_multi(
+            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("/{chat_history_id}", response_model=schemas.ChatHistory)
+def read_chat_history(
+    chat_history_id: int,
+    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=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")
+        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

@@ -5,4 +5,5 @@ from .company_crud import company
 from .subscription_crud import subscription
 from .document_crud import documents
 from .department_crud import department
-from .audit_crud import audit
+from .audit_crud import audit
+from .chat_crud import chat, chat_item

private_gpt/users/crud/chat_crud.py

@@ -0,0 +1,21 @@
+from typing import Optional, List, Union, Dict, Any
+from fastapi import HTTPException, status
+from sqlalchemy.orm import Session
+from sqlalchemy.exc import IntegrityError
+
+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: int) -> Optional[ChatHistory]:
+        return db.query(self.model).filter(ChatHistory.conversation_id == id).first()
+
+
+class CRUDChatItem(CRUDBase[ChatItem, ChatItemCreate, ChatItemUpdate]):
+    pass
+
+
+chat = CRUDChat(ChatHistory)
+chat_item = CRUDChatItem(ChatItem)

private_gpt/users/models/__init__.py

@@ -6,4 +6,5 @@ from .document import Document
 from .subscription import Subscription
 from .department import Department
 from .audit import Audit
-from .document_department import document_department_association
+from .document_department import document_department_association
+from .chat import ChatHistory, ChatItem

private_gpt/users/models/chat.py

@@ -0,0 +1,60 @@
+from datetime import datetime
+from sqlalchemy.orm import relationship
+from sqlalchemy import Column, Integer, String, DateTime, ForeignKey, Text, Boolean
+from private_gpt.users.db.base_class import Base
+
+
+class ChatHistory(Base):
+    """Models a chat history table"""
+
+    __tablename__ = "chat_history"
+
+    conversation_id = Column(Integer, nullable=False, primary_key=True)
+    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")
+
+    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 = "Untitled Chat"
+
+    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(Text, 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(Integer, 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}>"

private_gpt/users/models/user.py

@@ -51,6 +51,7 @@ class User(Base):
         Integer, ForeignKey("departments.id"), nullable=False)
     
     department = relationship("Department", back_populates="users")
+    chat_histories = relationship("ChatHistory", back_populates="user")
 
     __table_args__ = (
         UniqueConstraint('username', name='unique_username_no_spacing'),

private_gpt/users/schemas/__init__.py

@@ -1,9 +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 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
+from pydantic import BaseModel
+
+
+class ChatItemBase(BaseModel):
+    conversation_id: int
+    sender: str
+    content: Optional[str]
+
+
+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: int
+
+class ChatHistory(ChatHistoryBase):
+    conversation_id: int
+    created_at: datetime
+    updated_at: datetime
+    chat_items: List[ChatItem]
+
+    class Config:
+        orm_mode = True
+
+class ChatDelete(BaseModel):
+    conversation_id: int
+
+class CreateChatHistory(BaseModel):
+    user_id: int

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