khoj 1.20.3__py3-none-any.whl → 1.20.5.dev10__py3-none-any.whl

khoj/processor/conversation/prompts.py

@@ -208,10 +208,12 @@ Construct search queries to retrieve relevant information to answer the user's q
 - Add as much context from the previous questions and answers as required into your search queries.
 - Break messages into multiple search queries when required to retrieve the relevant information.
 - Add date filters to your search queries from questions and answers when required to retrieve the relevant information.
+- When asked a meta, vague or random questions, search for a variety of broad topics to answer the user's question.
 - Share relevant search queries as a JSON list of strings. Do not say anything else.
 
 Current Date: {day_of_week}, {current_date}
 User's Location: {location}
+{username}
 
 Examples:
 Q: How was my trip to Cambodia?
@@ -238,6 +240,9 @@ Khoj: ["What kind of plants do I have?", "What issues do my plants have?"]
 Q: Who all did I meet here yesterday?
 Khoj: ["Met in {location} on {yesterday_date} dt>='{yesterday_date}' dt<'{current_date}'"]
 
+Q: Share some random, interesting experiences from this month
+Khoj: ["Exciting travel adventures from {current_month}", "Fun social events dt>='{current_month}-01' dt<'{current_date}'", "Intense emotional experiences in {current_month}"]
+
 Chat History:
 {chat_history}
 What searches will you perform to answer the following question, using the chat history as reference? Respond only with relevant search queries as a valid JSON list of strings.
@@ -254,10 +259,12 @@ Construct search queries to retrieve relevant information to answer the user's q
 - Add as much context from the previous questions and answers as required into your search queries.
 - Break messages into multiple search queries when required to retrieve the relevant information.
 - Add date filters to your search queries from questions and answers when required to retrieve the relevant information.
+- When asked a meta, vague or random questions, search for a variety of broad topics to answer the user's question.
 
 What searches will you perform to answer the users question? Respond with search queries as list of strings in a JSON object.
 Current Date: {day_of_week}, {current_date}
 User's Location: {location}
+{username}
 
 Q: How was my trip to Cambodia?
 Khoj: {{"queries": ["How was my trip to Cambodia?"]}}
@@ -279,6 +286,10 @@ Q: How many tennis balls fit in the back of a 2002 Honda Civic?
 Khoj: {{"queries": ["What is the size of a tennis ball?", "What is the trunk size of a 2002 Honda Civic?"]}}
 A: 1085 tennis balls will fit in the trunk of a Honda Civic
 
+Q: Share some random, interesting experiences from this month
+Khoj: {{"queries": ["Exciting travel adventures from {current_month}", "Fun social events dt>='{current_month}-01' dt<'{current_date}'", "Intense emotional experiences in {current_month}"]}}
+A: You had a great time at the local beach with your friends, attended a music concert and had a deep conversation with your friend, Khalid.
+
 Q: Is Bob older than Tom?
 Khoj: {{"queries": ["When was Bob born?", "What is Tom's age?"]}}
 A: Yes, Bob is older than Tom. As Bob was born on 1984-01-01 and Tom is 30 years old.
@@ -305,11 +316,13 @@ Construct search queries to retrieve relevant information to answer the user's q
 - Add as much context from the previous questions and answers as required into your search queries.
 - Break messages into multiple search queries when required to retrieve the relevant information.
 - Add date filters to your search queries from questions and answers when required to retrieve the relevant information.
+- When asked a meta, vague or random questions, search for a variety of broad topics to answer the user's question.
 
 What searches will you perform to answer the users question? Respond with a JSON object with the key "queries" mapping to a list of searches you would perform on the user's knowledge base. Just return the queries and nothing else.
 
 Current Date: {day_of_week}, {current_date}
 User's Location: {location}
+{username}
 
 Here are some examples of how you can construct search queries to answer the user's question:
 
@@ -328,6 +341,11 @@ A: I can help you live healthier and happier across work and personal life
 User: Who all did I meet here yesterday?
 Assistant: {{"queries": ["Met in {location} on {yesterday_date} dt>='{yesterday_date}' dt<'{current_date}'"]}}
 A: Yesterday's note mentions your visit to your local beach with Ram and Shyam.
+
+User: Share some random, interesting experiences from this month
+Assistant: {{"queries": ["Exciting travel adventures from {current_month}", "Fun social events dt>='{current_month}-01' dt<'{current_date}'", "Intense emotional experiences in {current_month}"]}}
+A: You had a great time at the local beach with your friends, attended a music concert and had a deep conversation with your friend, Khalid.
+
 """.strip()
 )
 
@@ -525,6 +543,7 @@ Which webpages will you need to read to answer the user's question?
 Provide web page links as a list of strings in a JSON object.
 Current Date: {current_date}
 User's Location: {location}
+{username}
 
 Here are some examples:
 History:
@@ -571,6 +590,7 @@ What Google searches, if any, will you need to perform to answer the user's ques
 Provide search queries as a list of strings in a JSON object. Do not wrap the json in a codeblock.
 Current Date: {current_date}
 User's Location: {location}
+{username}
 
 Here are some examples:
 History:

khoj/processor/embeddings.py

@@ -95,11 +95,13 @@ class CrossEncoderModel:
         model_name: str = "mixedbread-ai/mxbai-rerank-xsmall-v1",
         cross_encoder_inference_endpoint: str = None,
         cross_encoder_inference_endpoint_api_key: str = None,
+        model_kwargs: dict = {},
     ):
         self.model_name = model_name
-        self.cross_encoder_model = CrossEncoder(model_name=self.model_name, device=get_device())
         self.inference_endpoint = cross_encoder_inference_endpoint
         self.api_key = cross_encoder_inference_endpoint_api_key
+        self.model_kwargs = merge_dicts(model_kwargs, {"device": get_device()})
+        self.cross_encoder_model = CrossEncoder(model_name=self.model_name, **self.model_kwargs)
 
     def inference_server_enabled(self) -> bool:
         return self.api_key is not None and self.inference_endpoint is not None

khoj/processor/tools/online_search.py

@@ -10,6 +10,7 @@ import aiohttp
 from bs4 import BeautifulSoup
 from markdownify import markdownify
 
+from khoj.database.models import KhojUser
 from khoj.routers.helpers import (
     ChatEvent,
     extract_relevant_info,
@@ -51,6 +52,7 @@ async def search_online(
     query: str,
     conversation_history: dict,
     location: LocationData,
+    user: KhojUser,
     send_status_func: Optional[Callable] = None,
     custom_filters: List[str] = [],
 ):
@@ -61,7 +63,7 @@ async def search_online(
         return
 
     # Breakdown the query into subqueries to get the correct answer
-    subqueries = await generate_online_subqueries(query, conversation_history, location)
+    subqueries = await generate_online_subqueries(query, conversation_history, location, user)
     response_dict = {}
 
     if subqueries:
@@ -126,14 +128,18 @@ async def search_with_google(query: str) -> Tuple[str, Dict[str, List[Dict]]]:
 
 
 async def read_webpages(
-    query: str, conversation_history: dict, location: LocationData, send_status_func: Optional[Callable] = None
+    query: str,
+    conversation_history: dict,
+    location: LocationData,
+    user: KhojUser,
+    send_status_func: Optional[Callable] = None,
 ):
     "Infer web pages to read from the query and extract relevant information from them"
     logger.info(f"Inferring web pages to read")
     if send_status_func:
         async for event in send_status_func(f"**Inferring web pages to read**"):
             yield {ChatEvent.STATUS: event}
-    urls = await infer_webpage_urls(query, conversation_history, location)
+    urls = await infer_webpage_urls(query, conversation_history, location, user)
 
     logger.info(f"Reading web pages at: {urls}")
     if send_status_func:

khoj/routers/api.py

@@ -388,6 +388,7 @@ async def extract_references_and_questions(
                 conversation_log=meta_log,
                 should_extract_questions=True,
                 location_data=location_data,
+                user=user,
                 max_prompt_size=conversation_config.max_prompt_size,
             )
         elif conversation_config.model_type == ChatModelOptions.ModelType.OPENAI:
@@ -402,7 +403,7 @@ async def extract_references_and_questions(
                 api_base_url=base_url,
                 conversation_log=meta_log,
                 location_data=location_data,
-                max_tokens=conversation_config.max_prompt_size,
+                user=user,
             )
         elif conversation_config.model_type == ChatModelOptions.ModelType.ANTHROPIC:
             api_key = conversation_config.openai_config.api_key
@@ -413,6 +414,7 @@ async def extract_references_and_questions(
                 api_key=api_key,
                 conversation_log=meta_log,
                 location_data=location_data,
+                user=user,
             )
 
     # Collate search results as context for GPT
@@ -545,15 +547,19 @@ async def post_automation(
     if not subject:
         subject = await acreate_title_from_query(q)
 
+    title = f"Automation: {subject}"
+
     # Create new Conversation Session associated with this new task
-    conversation = await ConversationAdapters.acreate_conversation_session(user, request.user.client_app)
+    conversation = await ConversationAdapters.acreate_conversation_session(user, request.user.client_app, title=title)
 
-    calling_url = request.url.replace(query=f"{request.url.query}&conversation_id={conversation.id}")
+    calling_url = request.url.replace(query=f"{request.url.query}")
 
     # Schedule automation with query_to_run, timezone, subject directly provided by user
     try:
         # Use the query to run as the scheduling request if the scheduling request is unset
-        automation = await schedule_automation(query_to_run, subject, crontime, timezone, q, user, calling_url)
+        automation = await schedule_automation(
+            query_to_run, subject, crontime, timezone, q, user, calling_url, conversation.id
+        )
     except Exception as e:
         logger.error(f"Error creating automation {q} for {user.email}: {e}", exc_info=True)
         return Response(
@@ -649,6 +655,16 @@ def edit_job(
     automation_metadata["query_to_run"] = query_to_run
     automation_metadata["subject"] = subject.strip()
     automation_metadata["crontime"] = crontime
+    conversation_id = automation_metadata.get("conversation_id")
+
+    if not conversation_id:
+        title = f"Automation: {subject}"
+
+        # Create new Conversation Session associated with this new task
+        conversation = ConversationAdapters.create_conversation_session(user, request.user.client_app, title=title)
+
+        conversation_id = conversation.id
+        automation_metadata["conversation_id"] = conversation_id
 
     # Modify automation with updated query, subject
     automation.modify(
@@ -659,6 +675,7 @@ def edit_job(
             "scheduling_request": q,
             "user": user,
             "calling_url": request.url,
+            "conversation_id": conversation_id,
         },
     )
 

khoj/routers/api_chat.py

@@ -146,7 +146,7 @@ async def sendfeedback(request: Request, data: FeedbackData):
 
 
 @api_chat.post("/speech")
-@requires(["authenticated", "premium"])
+@requires(["authenticated"])
 async def text_to_speech(
     request: Request,
     common: CommonQueryParams,
@@ -792,7 +792,7 @@ async def chat(
         if ConversationCommand.Online in conversation_commands:
             try:
                 async for result in search_online(
-                    defiltered_query, meta_log, location, partial(send_event, ChatEvent.STATUS), custom_filters
+                    defiltered_query, meta_log, location, user, partial(send_event, ChatEvent.STATUS), custom_filters
                 ):
                     if isinstance(result, dict) and ChatEvent.STATUS in result:
                         yield result[ChatEvent.STATUS]
@@ -809,7 +809,7 @@ async def chat(
         if ConversationCommand.Webpage in conversation_commands:
             try:
                 async for result in read_webpages(
-                    defiltered_query, meta_log, location, partial(send_event, ChatEvent.STATUS)
+                    defiltered_query, meta_log, location, user, partial(send_event, ChatEvent.STATUS)
                 ):
                     if isinstance(result, dict) and ChatEvent.STATUS in result:
                         yield result[ChatEvent.STATUS]

khoj/routers/email.py

@@ -117,7 +117,8 @@ def send_task_email(name, email, query, result, subject, is_image=False):
     template = env.get_template("task.html")
 
     if is_image:
-        result = f"![{subject}]({result})"
+        image = result.get("image")
+        result = f"![{subject}]({image})"
 
     html_result = markdown_it.MarkdownIt().render(result)
     html_content = template.render(name=name, subject=subject, query=query, result=html_result)

khoj/routers/helpers.py

@@ -340,11 +340,14 @@ async def aget_relevant_output_modes(query: str, conversation_history: dict, is_
         return ConversationCommand.Text
 
 
-async def infer_webpage_urls(q: str, conversation_history: dict, location_data: LocationData) -> List[str]:
+async def infer_webpage_urls(
+    q: str, conversation_history: dict, location_data: LocationData, user: KhojUser
+) -> List[str]:
     """
     Infer webpage links from the given query
     """
     location = f"{location_data.city}, {location_data.region}, {location_data.country}" if location_data else "Unknown"
+    username = prompts.user_name.format(name=user.get_full_name()) if user.get_full_name() else ""
     chat_history = construct_chat_history(conversation_history)
 
     utc_date = datetime.utcnow().strftime("%Y-%m-%d")
@@ -353,6 +356,7 @@ async def infer_webpage_urls(q: str, conversation_history: dict, location_data:
         query=q,
         chat_history=chat_history,
         location=location,
+        username=username,
     )
 
     with timer("Chat actor: Infer webpage urls to read", logger):
@@ -370,11 +374,14 @@ async def infer_webpage_urls(q: str, conversation_history: dict, location_data:
         raise ValueError(f"Invalid list of urls: {response}")
 
 
-async def generate_online_subqueries(q: str, conversation_history: dict, location_data: LocationData) -> List[str]:
+async def generate_online_subqueries(
+    q: str, conversation_history: dict, location_data: LocationData, user: KhojUser
+) -> List[str]:
     """
     Generate subqueries from the given query
     """
     location = f"{location_data.city}, {location_data.region}, {location_data.country}" if location_data else "Unknown"
+    username = prompts.user_name.format(name=user.get_full_name()) if user.get_full_name() else ""
     chat_history = construct_chat_history(conversation_history)
 
     utc_date = datetime.utcnow().strftime("%Y-%m-%d")
@@ -383,6 +390,7 @@ async def generate_online_subqueries(q: str, conversation_history: dict, locatio
         query=q,
         chat_history=chat_history,
         location=location,
+        username=username,
     )
 
     with timer("Chat actor: Generate online search subqueries", logger):
@@ -1074,7 +1082,13 @@ def should_notify(original_query: str, executed_query: str, ai_response: str) ->
 
 
 def scheduled_chat(
-    query_to_run: str, scheduling_request: str, subject: str, user: KhojUser, calling_url: URL, job_id: str = None
+    query_to_run: str,
+    scheduling_request: str,
+    subject: str,
+    user: KhojUser,
+    calling_url: URL,
+    job_id: str = None,
+    conversation_id: int = None,
 ):
     logger.info(f"Processing scheduled_chat: {query_to_run}")
     if job_id:
@@ -1101,6 +1115,10 @@ def scheduled_chat(
     # Replace the original scheduling query with the scheduled query
     query_dict["q"] = [query_to_run]
 
+    # Replace the original conversation_id with the conversation_id
+    if conversation_id:
+        query_dict["conversation_id"] = [conversation_id]
+
     # Construct the URL to call the chat API with the scheduled query string
     encoded_query = urlencode(query_dict, doseq=True)
     url = f"{scheme}://{calling_url.netloc}/api/chat?{encoded_query}"
@@ -1130,7 +1148,9 @@ def scheduled_chat(
     if raw_response.headers.get("Content-Type") == "application/json":
         response_map = raw_response.json()
         ai_response = response_map.get("response") or response_map.get("image")
-        is_image = response_map.get("image") is not None
+        is_image = False
+        if type(ai_response) == dict:
+            is_image = ai_response.get("image") is not None
     else:
         ai_response = raw_response.text
 
@@ -1142,9 +1162,11 @@ def scheduled_chat(
             return raw_response
 
 
-async def create_automation(q: str, timezone: str, user: KhojUser, calling_url: URL, meta_log: dict = {}):
+async def create_automation(
+    q: str, timezone: str, user: KhojUser, calling_url: URL, meta_log: dict = {}, conversation_id: int = None
+):
     crontime, query_to_run, subject = await schedule_query(q, meta_log)
-    job = await schedule_automation(query_to_run, subject, crontime, timezone, q, user, calling_url)
+    job = await schedule_automation(query_to_run, subject, crontime, timezone, q, user, calling_url, conversation_id)
     return job, crontime, query_to_run, subject
 
 
@@ -1156,6 +1178,7 @@ async def schedule_automation(
     scheduling_request: str,
     user: KhojUser,
     calling_url: URL,
+    conversation_id: int,
 ):
     # Disable minute level automation recurrence
     minute_value = crontime.split(" ")[0]
@@ -1173,6 +1196,7 @@ async def schedule_automation(
             "scheduling_request": scheduling_request,
             "subject": subject,
             "crontime": crontime,
+            "conversation_id": conversation_id,
         }
     )
     query_id = hashlib.md5(f"{query_to_run}_{crontime}".encode("utf-8")).hexdigest()
@@ -1191,6 +1215,7 @@ async def schedule_automation(
             "user": user,
             "calling_url": calling_url,
             "job_id": job_id,
+            "conversation_id": conversation_id,
         },
         id=job_id,
         name=job_metadata,

khoj/search_filter/file_filter.py

@@ -11,7 +11,8 @@ logger = logging.getLogger(__name__)
 
 
 class FileFilter(BaseFilter):
-    file_filter_regex = r'file:"(.+?)" ?'
+    file_filter_regex = r'(?<!-)file:"(.+?)" ?'
+    excluded_file_filter_regex = r'-file:"(.+?)" ?'
 
     def __init__(self, entry_key="file"):
         self.entry_key = entry_key
@@ -20,7 +21,9 @@ class FileFilter(BaseFilter):
 
     def get_filter_terms(self, query: str) -> List[str]:
         "Get all filter terms in query"
-        return [f"{self.convert_to_regex(term)}" for term in re.findall(self.file_filter_regex, query)]
+        required_files = [f"{required_file}" for required_file in re.findall(self.file_filter_regex, query)]
+        excluded_files = [f"-{excluded_file}" for excluded_file in re.findall(self.excluded_file_filter_regex, query)]
+        return required_files + excluded_files
 
     def convert_to_regex(self, file_filter: str) -> str:
         "Convert file filter to regex"

{khoj-1.20.3.dist-info → khoj-1.20.5.dev10.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: khoj
-Version: 1.20.3
+Version: 1.20.5.dev10
 Summary: Your Second Brain
 Project-URL: Homepage, https://khoj.dev
 Project-URL: Documentation, https://docs.khoj.dev