khoj 1.28.4.dev23__py3-none-any.whl → 1.28.4.dev77__py3-none-any.whl

khoj/routers/helpers.py

@@ -105,6 +105,7 @@ from khoj.utils.config import OfflineChatProcessorModel
 from khoj.utils.helpers import (
     LRU,
     ConversationCommand,
+    get_file_type,
     is_none_or_empty,
     is_valid_url,
     log_telemetry,
@@ -112,7 +113,7 @@ from khoj.utils.helpers import (
     timer,
     tool_descriptions_for_llm,
 )
-from khoj.utils.rawconfig import LocationData
+from khoj.utils.rawconfig import ChatRequestBody, FileAttachment, FileData, LocationData
 
 logger = logging.getLogger(__name__)
 
@@ -168,6 +169,12 @@ async def is_ready_to_chat(user: KhojUser):
     raise HTTPException(status_code=500, detail="Set your OpenAI API key or enable Local LLM via Khoj settings.")
 
 
+def get_file_content(file: UploadFile):
+    file_content = file.file.read()
+    file_type, encoding = get_file_type(file.content_type, file_content)
+    return FileData(name=file.filename, content=file_content, file_type=file_type, encoding=encoding)
+
+
 def update_telemetry_state(
     request: Request,
     telemetry_type: str,
@@ -249,6 +256,39 @@ async def agenerate_chat_response(*args):
     return await loop.run_in_executor(executor, generate_chat_response, *args)
 
 
+def gather_raw_query_files(
+    query_files: Dict[str, str],
+):
+    """
+    Gather contextual data from the given (raw) files
+    """
+
+    if len(query_files) == 0:
+        return ""
+
+    contextual_data = " ".join(
+        [f"File: {file_name}\n\n{file_content}\n\n" for file_name, file_content in query_files.items()]
+    )
+    return f"I have attached the following files:\n\n{contextual_data}"
+
+
+async def acreate_title_from_history(
+    user: KhojUser,
+    conversation: Conversation,
+):
+    """
+    Create a title from the given conversation history
+    """
+    chat_history = construct_chat_history(conversation.conversation_log)
+
+    title_generation_prompt = prompts.conversation_title_generation.format(chat_history=chat_history)
+
+    with timer("Chat actor: Generate title from conversation history", logger):
+        response = await send_message_to_model_wrapper(title_generation_prompt, user=user)
+
+    return response.strip()
+
+
 async def acreate_title_from_query(query: str, user: KhojUser = None) -> str:
     """
     Create a title from the given query
@@ -261,11 +301,15 @@ async def acreate_title_from_query(query: str, user: KhojUser = None) -> str:
     return response.strip()
 
 
-async def acheck_if_safe_prompt(system_prompt: str, user: KhojUser = None) -> Tuple[bool, str]:
+async def acheck_if_safe_prompt(system_prompt: str, user: KhojUser = None, lax: bool = False) -> Tuple[bool, str]:
     """
     Check if the system prompt is safe to use
     """
-    safe_prompt_check = prompts.personality_prompt_safety_expert.format(prompt=system_prompt)
+    safe_prompt_check = (
+        prompts.personality_prompt_safety_expert.format(prompt=system_prompt)
+        if not lax
+        else prompts.personality_prompt_safety_expert_lax.format(prompt=system_prompt)
+    )
     is_safe = True
     reason = ""
 
@@ -294,6 +338,7 @@ async def aget_relevant_information_sources(
     user: KhojUser,
     query_images: List[str] = None,
     agent: Agent = None,
+    query_files: str = None,
     tracer: dict = {},
 ):
     """
@@ -331,6 +376,7 @@ async def aget_relevant_information_sources(
             relevant_tools_prompt,
             response_type="json_object",
             user=user,
+            query_files=query_files,
             tracer=tracer,
         )
 
@@ -440,6 +486,7 @@ async def infer_webpage_urls(
     user: KhojUser,
     query_images: List[str] = None,
     agent: Agent = None,
+    query_files: str = None,
     tracer: dict = {},
 ) -> List[str]:
     """
@@ -469,6 +516,7 @@ async def infer_webpage_urls(
             query_images=query_images,
             response_type="json_object",
             user=user,
+            query_files=query_files,
             tracer=tracer,
         )
 
@@ -494,6 +542,7 @@ async def generate_online_subqueries(
     user: KhojUser,
     query_images: List[str] = None,
     agent: Agent = None,
+    query_files: str = None,
     tracer: dict = {},
 ) -> Set[str]:
     """
@@ -523,6 +572,7 @@ async def generate_online_subqueries(
             query_images=query_images,
             response_type="json_object",
             user=user,
+            query_files=query_files,
             tracer=tracer,
         )
 
@@ -645,26 +695,38 @@ async def generate_summary_from_files(
     query_images: List[str] = None,
     agent: Agent = None,
     send_status_func: Optional[Callable] = None,
+    query_files: str = None,
     tracer: dict = {},
 ):
     try:
-        file_object = None
+        file_objects = None
         if await EntryAdapters.aagent_has_entries(agent):
             file_names = await EntryAdapters.aget_agent_entry_filepaths(agent)
             if len(file_names) > 0:
-                file_object = await FileObjectAdapters.async_get_file_objects_by_name(None, file_names.pop(), agent)
-
-        if len(file_filters) > 0:
-            file_object = await FileObjectAdapters.async_get_file_objects_by_name(user, file_filters[0])
+                file_objects = await FileObjectAdapters.aget_file_objects_by_name(None, file_names.pop(), agent)
 
-        if len(file_object) == 0:
-            response_log = "Sorry, I couldn't find the full text of this file."
+        if (file_objects and len(file_objects) == 0 and not query_files) or (not file_objects and not query_files):
+            response_log = "Sorry, I couldn't find anything to summarize."
             yield response_log
             return
-        contextual_data = " ".join([file.raw_text for file in file_object])
+
+        contextual_data = " ".join([f"File: {file.file_name}\n\n{file.raw_text}" for file in file_objects])
+
+        if query_files:
+            contextual_data += f"\n\n{query_files}"
+
         if not q:
             q = "Create a general summary of the file"
-        async for result in send_status_func(f"**Constructing Summary Using:** {file_object[0].file_name}"):
+
+        file_names = [file.file_name for file in file_objects]
+        file_names.extend(file_filters)
+
+        all_file_names = ""
+
+        for file_name in file_names:
+            all_file_names += f"- {file_name}\n"
+
+        async for result in send_status_func(f"**Constructing Summary Using:**\n{all_file_names}"):
             yield {ChatEvent.STATUS: result}
 
         response = await extract_relevant_summary(
@@ -694,6 +756,7 @@ async def generate_excalidraw_diagram(
     user: KhojUser = None,
     agent: Agent = None,
     send_status_func: Optional[Callable] = None,
+    query_files: str = None,
     tracer: dict = {},
 ):
     if send_status_func:
@@ -709,6 +772,7 @@ async def generate_excalidraw_diagram(
         query_images=query_images,
         user=user,
         agent=agent,
+        query_files=query_files,
         tracer=tracer,
     )
 
@@ -735,6 +799,7 @@ async def generate_better_diagram_description(
     query_images: List[str] = None,
     user: KhojUser = None,
     agent: Agent = None,
+    query_files: str = None,
     tracer: dict = {},
 ) -> str:
     """
@@ -773,7 +838,11 @@ async def generate_better_diagram_description(
 
     with timer("Chat actor: Generate better diagram description", logger):
         response = await send_message_to_model_wrapper(
-            improve_diagram_description_prompt, query_images=query_images, user=user, tracer=tracer
+            improve_diagram_description_prompt,
+            query_images=query_images,
+            user=user,
+            query_files=query_files,
+            tracer=tracer,
         )
         response = response.strip()
         if response.startswith(('"', "'")) and response.endswith(('"', "'")):
@@ -820,6 +889,7 @@ async def generate_better_image_prompt(
     query_images: Optional[List[str]] = None,
     user: KhojUser = None,
     agent: Agent = None,
+    query_files: str = "",
     tracer: dict = {},
 ) -> str:
     """
@@ -868,7 +938,7 @@ async def generate_better_image_prompt(
 
     with timer("Chat actor: Generate contextual image prompt", logger):
         response = await send_message_to_model_wrapper(
-            image_prompt, query_images=query_images, user=user, tracer=tracer
+            image_prompt, query_images=query_images, user=user, query_files=query_files, tracer=tracer
         )
         response = response.strip()
         if response.startswith(('"', "'")) and response.endswith(('"', "'")):
@@ -884,6 +954,7 @@ async def send_message_to_model_wrapper(
     user: KhojUser = None,
     query_images: List[str] = None,
     context: str = "",
+    query_files: str = None,
     tracer: dict = {},
 ):
     conversation_config: ChatModelOptions = await ConversationAdapters.aget_default_conversation_config(user)
@@ -923,6 +994,7 @@ async def send_message_to_model_wrapper(
             max_prompt_size=max_tokens,
             vision_enabled=vision_available,
             model_type=conversation_config.model_type,
+            query_files=query_files,
         )
 
         return send_message_to_model_offline(
@@ -949,6 +1021,7 @@ async def send_message_to_model_wrapper(
             vision_enabled=vision_available,
             query_images=query_images,
             model_type=conversation_config.model_type,
+            query_files=query_files,
         )
 
         return send_message_to_model(
@@ -971,6 +1044,7 @@ async def send_message_to_model_wrapper(
             vision_enabled=vision_available,
             query_images=query_images,
             model_type=conversation_config.model_type,
+            query_files=query_files,
         )
 
         return anthropic_send_message_to_model(
@@ -992,6 +1066,7 @@ async def send_message_to_model_wrapper(
             vision_enabled=vision_available,
             query_images=query_images,
             model_type=conversation_config.model_type,
+            query_files=query_files,
         )
 
         return gemini_send_message_to_model(
@@ -1006,6 +1081,7 @@ def send_message_to_model_wrapper_sync(
     system_message: str = "",
     response_type: str = "text",
     user: KhojUser = None,
+    query_files: str = "",
     tracer: dict = {},
 ):
     conversation_config: ChatModelOptions = ConversationAdapters.get_default_conversation_config(user)
@@ -1030,6 +1106,7 @@ def send_message_to_model_wrapper_sync(
             max_prompt_size=max_tokens,
             vision_enabled=vision_available,
             model_type=conversation_config.model_type,
+            query_files=query_files,
         )
 
         return send_message_to_model_offline(
@@ -1051,6 +1128,7 @@ def send_message_to_model_wrapper_sync(
             max_prompt_size=max_tokens,
             vision_enabled=vision_available,
             model_type=conversation_config.model_type,
+            query_files=query_files,
         )
 
         openai_response = send_message_to_model(
@@ -1072,6 +1150,7 @@ def send_message_to_model_wrapper_sync(
             max_prompt_size=max_tokens,
             vision_enabled=vision_available,
             model_type=conversation_config.model_type,
+            query_files=query_files,
         )
 
         return anthropic_send_message_to_model(
@@ -1091,6 +1170,7 @@ def send_message_to_model_wrapper_sync(
             max_prompt_size=max_tokens,
             vision_enabled=vision_available,
             model_type=conversation_config.model_type,
+            query_files=query_files,
         )
 
         return gemini_send_message_to_model(
@@ -1120,8 +1200,10 @@ def generate_chat_response(
     user_name: Optional[str] = None,
     meta_research: str = "",
     query_images: Optional[List[str]] = None,
-    tracer: dict = {},
     train_of_thought: List[Any] = [],
+    query_files: str = None,
+    raw_query_files: List[FileAttachment] = None,
+    tracer: dict = {},
 ) -> Tuple[Union[ThreadedGenerator, Iterator[str]], Dict[str, str]]:
     # Initialize Variables
     chat_response = None
@@ -1142,8 +1224,9 @@ def generate_chat_response(
             client_application=client_application,
             conversation_id=conversation_id,
             query_images=query_images,
-            tracer=tracer,
             train_of_thought=train_of_thought,
+            raw_query_files=raw_query_files,
+            tracer=tracer,
         )
 
         query_to_run = q
@@ -1177,6 +1260,7 @@ def generate_chat_response(
                 location_data=location_data,
                 user_name=user_name,
                 agent=agent,
+                query_files=query_files,
                 tracer=tracer,
             )
 
@@ -1202,6 +1286,7 @@ def generate_chat_response(
                 user_name=user_name,
                 agent=agent,
                 vision_available=vision_available,
+                query_files=query_files,
                 tracer=tracer,
             )
 
@@ -1224,6 +1309,7 @@ def generate_chat_response(
                 user_name=user_name,
                 agent=agent,
                 vision_available=vision_available,
+                query_files=query_files,
                 tracer=tracer,
             )
         elif conversation_config.model_type == ChatModelOptions.ModelType.GOOGLE:
@@ -1243,7 +1329,9 @@ def generate_chat_response(
                 location_data=location_data,
                 user_name=user_name,
                 agent=agent,
+                query_images=query_images,
                 vision_available=vision_available,
+                query_files=query_files,
                 tracer=tracer,
             )
 
@@ -1256,23 +1344,6 @@ def generate_chat_response(
     return chat_response, metadata
 
 
-class ChatRequestBody(BaseModel):
-    q: str
-    n: Optional[int] = 7
-    d: Optional[float] = None
-    stream: Optional[bool] = False
-    title: Optional[str] = None
-    conversation_id: Optional[str] = None
-    turn_id: Optional[str] = None
-    city: Optional[str] = None
-    region: Optional[str] = None
-    country: Optional[str] = None
-    country_code: Optional[str] = None
-    timezone: Optional[str] = None
-    images: Optional[list[str]] = None
-    create_new: Optional[bool] = False
-
-
 class DeleteMessageRequestBody(BaseModel):
     conversation_id: str
     turn_id: str
@@ -1904,10 +1975,10 @@ def get_user_config(user: KhojUser, request: Request, is_detailed: bool = False)
 
 
 def configure_content(
+    user: KhojUser,
     files: Optional[dict[str, dict[str, str]]],
     regenerate: bool = False,
     t: Optional[state.SearchType] = state.SearchType.All,
-    user: KhojUser = None,
 ) -> bool:
     success = True
     if t == None:

khoj/routers/notion.py

@@ -80,6 +80,6 @@ async def notion_auth_callback(request: Request, background_tasks: BackgroundTas
     notion_redirect = str(request.app.url_path_for("config_page"))
 
     # Trigger an async job to configure_content. Let it run without blocking the response.
-    background_tasks.add_task(run_in_executor, configure_content, {}, False, SearchType.Notion, user)
+    background_tasks.add_task(run_in_executor, configure_content, user, {}, False, SearchType.Notion)
 
     return RedirectResponse(notion_redirect)

khoj/routers/research.py

@@ -11,6 +11,7 @@ from khoj.processor.conversation import prompts
 from khoj.processor.conversation.utils import (
     InformationCollectionIteration,
     clean_json,
+    construct_chat_history,
     construct_iteration_history,
     construct_tool_chat_history,
 )
@@ -19,8 +20,6 @@ from khoj.processor.tools.run_code import run_code
 from khoj.routers.api import extract_references_and_questions
 from khoj.routers.helpers import (
     ChatEvent,
-    construct_chat_history,
-    extract_relevant_info,
     generate_summary_from_files,
     send_message_to_model_wrapper,
 )
@@ -47,6 +46,7 @@ async def apick_next_tool(
     max_iterations: int = 5,
     send_status_func: Optional[Callable] = None,
     tracer: dict = {},
+    query_files: str = None,
 ):
     """Given a query, determine which of the available tools the agent should use in order to answer appropriately."""
 
@@ -92,6 +92,7 @@ async def apick_next_tool(
                 response_type="json_object",
                 user=user,
                 query_images=query_images,
+                query_files=query_files,
                 tracer=tracer,
             )
     except Exception as e:
@@ -151,6 +152,7 @@ async def execute_information_collection(
     location: LocationData = None,
     file_filters: List[str] = [],
     tracer: dict = {},
+    query_files: str = None,
 ):
     current_iteration = 0
     MAX_ITERATIONS = 5
@@ -174,6 +176,7 @@ async def execute_information_collection(
             MAX_ITERATIONS,
             send_status_func,
             tracer=tracer,
+            query_files=query_files,
         ):
             if isinstance(result, dict) and ChatEvent.STATUS in result:
                 yield result[ChatEvent.STATUS]
@@ -204,6 +207,7 @@ async def execute_information_collection(
                 previous_inferred_queries=previous_inferred_queries,
                 agent=agent,
                 tracer=tracer,
+                query_files=query_files,
             ):
                 if isinstance(result, dict) and ChatEvent.STATUS in result:
                     yield result[ChatEvent.STATUS]
@@ -265,6 +269,7 @@ async def execute_information_collection(
                     query_images=query_images,
                     agent=agent,
                     tracer=tracer,
+                    query_files=query_files,
                 ):
                     if isinstance(result, dict) and ChatEvent.STATUS in result:
                         yield result[ChatEvent.STATUS]
@@ -295,6 +300,7 @@ async def execute_information_collection(
                     send_status_func,
                     query_images=query_images,
                     agent=agent,
+                    query_files=query_files,
                     tracer=tracer,
                 ):
                     if isinstance(result, dict) and ChatEvent.STATUS in result:
@@ -320,6 +326,7 @@ async def execute_information_collection(
                     query_images=query_images,
                     agent=agent,
                     send_status_func=send_status_func,
+                    query_files=query_files,
                 ):
                     if isinstance(result, dict) and ChatEvent.STATUS in result:
                         yield result[ChatEvent.STATUS]

khoj/search_type/text_search.py

@@ -208,7 +208,7 @@ def setup(
     text_to_entries: Type[TextToEntries],
     files: dict[str, str],
     regenerate: bool,
-    user: KhojUser = None,
+    user: KhojUser,
     config=None,
 ) -> Tuple[int, int]:
     if config:

khoj/utils/fs_syncer.py

@@ -8,6 +8,7 @@ from bs4 import BeautifulSoup
 from magika import Magika
 
 from khoj.database.models import (
+    KhojUser,
     LocalMarkdownConfig,
     LocalOrgConfig,
     LocalPdfConfig,
@@ -21,7 +22,7 @@ logger = logging.getLogger(__name__)
 magika = Magika()
 
 
-def collect_files(search_type: Optional[SearchType] = SearchType.All, user=None) -> dict:
+def collect_files(user: KhojUser, search_type: Optional[SearchType] = SearchType.All) -> dict:
     files: dict[str, dict] = {"docx": {}, "image": {}}
 
     if search_type == SearchType.All or search_type == SearchType.Org:

khoj/utils/rawconfig.py

@@ -138,6 +138,38 @@ class SearchResponse(ConfigBase):
     corpus_id: str
 
 
+class FileData(BaseModel):
+    name: str
+    content: bytes
+    file_type: str
+    encoding: str | None = None
+
+
+class FileAttachment(BaseModel):
+    name: str
+    content: str
+    file_type: str
+    size: int
+
+
+class ChatRequestBody(BaseModel):
+    q: str
+    n: Optional[int] = 7
+    d: Optional[float] = None
+    stream: Optional[bool] = False
+    title: Optional[str] = None
+    conversation_id: Optional[str] = None
+    turn_id: Optional[str] = None
+    city: Optional[str] = None
+    region: Optional[str] = None
+    country: Optional[str] = None
+    country_code: Optional[str] = None
+    timezone: Optional[str] = None
+    images: Optional[list[str]] = None
+    files: Optional[list[FileAttachment]] = []
+    create_new: Optional[bool] = False
+
+
 class Entry:
     raw: str
     compiled: str

{khoj-1.28.4.dev23.dist-info → khoj-1.28.4.dev77.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: khoj
-Version: 1.28.4.dev23
+Version: 1.28.4.dev77
 Summary: Your Second Brain
 Project-URL: Homepage, https://khoj.dev
 Project-URL: Documentation, https://docs.khoj.dev