log to huggingface

climateqa/chat.py

@@ -12,15 +12,11 @@ from .handle_stream_events import (
     convert_to_docs_to_html,
     stream_answer,
     handle_retrieved_owid_graphs,
-    serialize_docs,
 )
-
-# Function to log data on Azure
-def log_on_azure(file, logs, share_client):
-    logs = json.dumps(logs)
-    file_client = share_client.get_file_client(file)
-    file_client.upload_file(logs)
-
+from .logging import (
+    log_interaction_to_huggingface
+)
+        
 # Chat functions
 def start_chat(query, history, search_only):
     history = history + [ChatMessage(role="user", content=query)]
@@ -32,28 +28,6 @@ def start_chat(query, history, search_only):
 def finish_chat():
     return gr.update(interactive=True, value="")
 
-def log_interaction_to_azure(history, output_query, sources, docs, share_client, user_id):
-    try:
-        # Log interaction to Azure if not in local environment
-        if os.getenv("GRADIO_ENV") != "local":
-            timestamp = str(datetime.now().timestamp())
-            prompt = history[1]["content"]
-            logs = {
-                "user_id": str(user_id),
-                "prompt": prompt,
-                "query": prompt,
-                "question": output_query,
-                "sources": sources,
-                "docs": serialize_docs(docs),
-                "answer": history[-1].content,
-                "time": timestamp,
-            }
-            log_on_azure(f"{timestamp}.json", logs, share_client)
-    except Exception as e:
-        print(f"Error logging on Azure Blob Storage: {e}")
-        error_msg = f"ClimateQ&A Error: {str(e)[:100]} - The error has been noted, try another question and if the error remains, you can contact us :)"
-        raise gr.Error(error_msg)
-
 def handle_numerical_data(event):
     if event["name"] == "retrieve_drias_data" and event["event"] == "on_chain_end":
         numerical_data = event["data"]["output"]["drias_data"]
@@ -61,27 +35,6 @@ def handle_numerical_data(event):
         return numerical_data, sql_query
     return None, None
     
-def log_drias_interaction_to_azure(query, sql_query, data, share_client, user_id):
-    try:
-        # Log interaction to Azure if not in local environment
-        if os.getenv("GRADIO_ENV") != "local":
-            timestamp = str(datetime.now().timestamp())
-            logs = {
-                "user_id": str(user_id),
-                "query": query,
-                "sql_query": sql_query,
-                # "data": data.to_dict() if data is not None else None,
-                "time": timestamp,
-            }
-            log_on_azure(f"drias_{timestamp}.json", logs, share_client)
-            print(f"Logged Drias interaction to Azure Blob Storage: {logs}")
-        else:
-            print("share_client or user_id is None, or GRADIO_ENV is local")
-    except Exception as e:
-        print(f"Error logging Drias interaction on Azure Blob Storage: {e}")
-        error_msg = f"Drias Error: {str(e)[:100]} - The error has been noted, try another question and if the error remains, you can contact us :)"
-        raise gr.Error(error_msg)
-
 # Main chat function
 async def chat_stream(
     agent : CompiledStateGraph,
@@ -235,9 +188,7 @@ async def chat_stream(
         print(f"Event {event} has failed")
         raise gr.Error(str(e))
 
- 
-
     # Call the function to log interaction
-    log_interaction_to_azure(history, output_query, sources, docs, share_client, user_id)
+    log_interaction_to_huggingface(history, output_query, sources, docs, share_client, user_id)
 
     yield history, docs_html, output_query, output_language, related_contents, graphs_html, follow_up_examples#, vanna_data

climateqa/engine/talk_to_data/main.py

@@ -1,5 +1,6 @@
 from climateqa.engine.talk_to_data.workflow import drias_workflow
 from climateqa.engine.llm import get_llm
+from climateqa.logging import log_drias_interaction_to_huggingface
 import ast
 
 llm = get_llm(provider="openai")
@@ -37,7 +38,7 @@ def ask_llm_column_names(sql_query: str, llm) -> list[str]:
     columns_list = ast.literal_eval(columns.strip("```python\n").strip())
     return columns_list
 
-async def ask_drias(query: str, index_state: int = 0) -> tuple:
+async def ask_drias(query: str, index_state: int = 0, user_id: str = None) -> tuple:
     """Main function to process a DRIAS query and return results.
     
     This function orchestrates the DRIAS workflow, processing a user query to generate
@@ -85,6 +86,8 @@ async def ask_drias(query: str, index_state: int = 0) -> tuple:
     sql_query = sql_queries[index_state]
     dataframe = result_dataframes[index_state]
     figure = figures[index_state](dataframe)
+    
+    log_drias_interaction_to_huggingface(query, sql_query, user_id)
 
     return sql_query, dataframe, figure, sql_queries, result_dataframes, figures, index_state, table_list, ""
 

climateqa/handle_stream_events.py

@@ -1,7 +1,7 @@
 from langchain_core.runnables.schema import StreamEvent
 from gradio import ChatMessage
 from climateqa.engine.chains.prompts import audience_prompts
-from front.utils import make_html_source,parse_output_llm_with_sources,serialize_docs,make_toolbox,generate_html_graphs
+from front.utils import make_html_source,parse_output_llm_with_sources
 import numpy as np
 
 def init_audience(audience :str) -> str:

climateqa/logging.py

@@ -0,0 +1,194 @@
+import os
+from datetime import datetime
+import json
+from huggingface_hub import HfApi
+import gradio as gr
+import csv
+
+def serialize_docs(docs:list)->list:
+    new_docs = []
+    for doc in docs:
+        new_doc = {}
+        new_doc["page_content"] = doc.page_content
+        new_doc["metadata"] = doc.metadata
+        new_docs.append(new_doc)
+    return new_docs
+
+## AZURE LOGGING - DEPRECATED
+
+# def log_on_azure(file, logs, share_client):
+#     """Log data to Azure Blob Storage.
+    
+#     Args:
+#         file (str): Name of the file to store logs
+#         logs (dict): Log data to store
+#         share_client: Azure share client instance
+#     """
+#     logs = json.dumps(logs)
+#     file_client = share_client.get_file_client(file)
+#     file_client.upload_file(logs)
+
+
+# def log_interaction_to_azure(history, output_query, sources, docs, share_client, user_id):
+#     """Log chat interaction to Azure and Hugging Face.
+    
+#     Args:
+#         history (list): Chat message history
+#         output_query (str): Processed query
+#         sources (list): Knowledge base sources used
+#         docs (list): Retrieved documents
+#         share_client: Azure share client instance
+#         user_id (str): User identifier
+#     """
+#     try:
+#         # Log interaction to Azure if not in local environment
+#         if os.getenv("GRADIO_ENV") != "local":
+#             timestamp = str(datetime.now().timestamp())
+#             prompt = history[1]["content"]
+#             logs = {
+#                 "user_id": str(user_id),
+#                 "prompt": prompt,
+#                 "query": prompt,
+#                 "question": output_query,
+#                 "sources": sources,
+#                 "docs": serialize_docs(docs),
+#                 "answer": history[-1].content,
+#                 "time": timestamp,
+#             }
+#             # Log to Azure
+#             log_on_azure(f"{timestamp}.json", logs, share_client)
+#     except Exception as e:
+#         print(f"Error logging on Azure Blob Storage: {e}")
+#         error_msg = f"ClimateQ&A Error: {str(e)[:100]} - The error has been noted, try another question and if the error remains, you can contact us :)"
+#         raise gr.Error(error_msg)
+    
+# def log_drias_interaction_to_azure(query, sql_query, data, share_client, user_id):
+#     """Log Drias data interaction to Azure and Hugging Face.
+    
+#     Args:
+#         query (str): User query
+#         sql_query (str): SQL query used
+#         data: Retrieved data
+#         share_client: Azure share client instance
+#         user_id (str): User identifier
+#     """
+#     try:
+#         # Log interaction to Azure if not in local environment
+#         if os.getenv("GRADIO_ENV") != "local":
+#             timestamp = str(datetime.now().timestamp())
+#             logs = {
+#                 "user_id": str(user_id),
+#                 "query": query,
+#                 "sql_query": sql_query,
+#                 "time": timestamp,
+#             }
+#             log_on_azure(f"drias_{timestamp}.json", logs, share_client)
+#             print(f"Logged Drias interaction to Azure Blob Storage: {logs}")
+#         else:
+#             print("share_client or user_id is None, or GRADIO_ENV is local")
+#     except Exception as e:
+#         print(f"Error logging Drias interaction on Azure Blob Storage: {e}")
+#         error_msg = f"Drias Error: {str(e)[:100]} - The error has been noted, try another question and if the error remains, you can contact us :)"
+#         raise gr.Error(error_msg)    
+    
+## HUGGING FACE LOGGING
+
+def log_on_huggingface(log_filename, logs):
+    """Log data to Hugging Face dataset repository.
+    
+    Args:
+        log_filename (str): Name of the file to store logs
+        logs (dict): Log data to store
+    """
+    try:
+        # Get Hugging Face token from environment
+        hf_token = os.getenv("HF_LOGS_TOKEN")
+        if not hf_token:
+            print("HF_LOGS_TOKEN not found in environment variables")
+            return
+
+        # Get repository name from environment or use default
+        repo_id = os.getenv("HF_DATASET_REPO", "timeki/climateqa_logs")
+        
+        # Initialize HfApi
+        api = HfApi(token=hf_token)
+        
+        # Add timestamp to the log data
+        logs["timestamp"] = datetime.now().strftime("%Y%m%d_%H%M%S_%f")
+        
+        # Convert logs to JSON string
+        logs_json = json.dumps(logs)
+        
+        # Upload directly from memory
+        api.upload_file(
+            path_or_fileobj=logs_json.encode('utf-8'),
+            path_in_repo=log_filename,
+            repo_id=repo_id,
+            repo_type="dataset"
+        )
+            
+    except Exception as e:
+        print(f"Error logging to Hugging Face: {e}")
+
+    
+def log_interaction_to_huggingface(history, output_query, sources, docs, share_client, user_id):
+    """Log chat interaction to Hugging Face.
+    
+    Args:
+        history (list): Chat message history
+        output_query (str): Processed query
+        sources (list): Knowledge base sources used
+        docs (list): Retrieved documents
+        share_client: Azure share client instance (unused in this function)
+        user_id (str): User identifier
+    """
+    try:
+        # Log interaction if not in local environment
+        if os.getenv("GRADIO_ENV") != "local":
+            timestamp = str(datetime.now().timestamp())
+            prompt = history[1]["content"]
+            logs = {
+                "user_id": str(user_id),
+                "prompt": prompt,
+                "query": prompt,
+                "question": output_query,
+                "sources": sources,
+                "docs": serialize_docs(docs),
+                "answer": history[-1].content,
+                "time": timestamp,
+            }
+            # Log to Hugging Face
+            log_on_huggingface(f"chat/{timestamp}.json", logs)
+    except Exception as e:
+        print(f"Error logging to Hugging Face: {e}")
+        error_msg = f"ClimateQ&A Error: {str(e)[:100]} - The error has been noted, try another question and if the error remains, you can contact us :)"
+        raise gr.Error(error_msg)
+
+def log_drias_interaction_to_huggingface(query, sql_query, user_id):
+    """Log Drias data interaction to Hugging Face.
+    
+    Args:
+        query (str): User query
+        sql_query (str): SQL query used
+        data: Retrieved data
+        user_id (str): User identifier
+    """
+    try:
+        if os.getenv("GRADIO_ENV") != "local":
+            timestamp = str(datetime.now().timestamp())
+            logs = {
+                "user_id": str(user_id),
+                "query": query,
+                "sql_query": sql_query,
+                "time": timestamp,
+            }
+            log_on_huggingface(f"drias/drias_{timestamp}.json", logs)
+            print(f"Logged Drias interaction to Hugging Face: {logs}")
+        else:
+            print("share_client or user_id is None, or GRADIO_ENV is local")
+    except Exception as e:
+        print(f"Error logging Drias interaction to Hugging Face: {e}")
+        error_msg = f"Drias Error: {str(e)[:100]} - The error has been noted, try another question and if the error remains, you can contact us :)"
+        raise gr.Error(error_msg)
+
+

front/tabs/chat_interface.py

@@ -39,7 +39,7 @@ What do you want to learn ?
 # """
 
 init_prompt_poc = """
-Hello, I am ClimateQ&A, a conversational assistant designed to help you understand climate change and biodiversity loss. I will answer your questions by **sifting through the IPCC and IPBES scientific reports, the Paris Climate Action Plan (PCAET), the Biodiversity Plan 2018-2024, and the Acclimaterra reports from the Nouvelle-Aquitaine Region**.
+Hello, I am ClimateQ&A, a conversational assistant designed to help you understand climate change and biodiversity loss. I will answer your questions by **sifting through the IPCC and IPBES scientific reports, the Paris Climate Action Plan (PCAET), the Paris Biodiversity Plan 2018-2024, and the Acclimaterra reports from the Nouvelle-Aquitaine Region**.
 
 ❓ How to use
 - **Language**: You can ask me your questions in any language. 

front/tabs/tab_drias.py

@@ -5,8 +5,6 @@ import pandas as pd
 
 from climateqa.engine.talk_to_data.main import ask_drias
 from climateqa.engine.talk_to_data.config import DRIAS_MODELS, DRIAS_UI_TEXT
-from climateqa.chat import log_drias_interaction_to_azure
-
 
 class DriasUIElements(TypedDict):
     tab: gr.Tab
@@ -28,8 +26,8 @@ class DriasUIElements(TypedDict):
     next_button: gr.Button
 
 
-async def ask_drias_query(query: str, index_state: int):
-    result = await ask_drias(query, index_state)
+async def ask_drias_query(query: str, index_state: int, user_id: str):
+    result = await ask_drias(query, index_state, user_id)
     return result
 
 
@@ -196,19 +194,7 @@ def create_drias_ui() -> DriasUIElements:
             next_button=next_button
         )
 
-def log_drias_to_azure(query: str, sql_query: str, data, share_client, user_id):
-    """Log Drias interaction to Azure storage."""
-    print("log_drias_to_azure")
-    if share_client is not None and user_id is not None:
-        log_drias_interaction_to_azure(
-            query=query,
-            sql_query=sql_query,
-            data=data,
-            share_client=share_client,
-            user_id=user_id
-        )
-    else:
-        print("share_client or user_id is None")
+
 
 def setup_drias_events(ui_elements: DriasUIElements, share_client=None, user_id=None) -> None:
     """Set up all event handlers for the DRIAS tab."""
@@ -218,10 +204,7 @@ def setup_drias_events(ui_elements: DriasUIElements, share_client=None, user_id=
     plots_state = gr.State([])
     index_state = gr.State(0)
     table_names_list = gr.State([])
-    
-    def log_drias_interaction(query: str, sql_query: str, data: pd.DataFrame):
-        log_drias_to_azure(query, sql_query, data, share_client, user_id)
-    
+    user_id = gr.State(user_id)
 
     # Handle example selection
     ui_elements["examples_hidden"].change(
@@ -230,7 +213,7 @@ def setup_drias_events(ui_elements: DriasUIElements, share_client=None, user_id=
         outputs=[ui_elements["details_accordion"], ui_elements["drias_direct_question"]]
     ).then(
         ask_drias_query,
-        inputs=[ui_elements["examples_hidden"], index_state],
+        inputs=[ui_elements["examples_hidden"], index_state, user_id],
         outputs=[
             ui_elements["drias_sql_query"],
             ui_elements["drias_table"],
@@ -242,10 +225,6 @@ def setup_drias_events(ui_elements: DriasUIElements, share_client=None, user_id=
             table_names_list,
             ui_elements["result_text"],
         ],
-    ).then(
-        log_drias_interaction,
-        inputs=[ui_elements["examples_hidden"], ui_elements["drias_sql_query"], ui_elements["drias_table"]],
-        outputs=[],
     ).then(
         show_results,
         inputs=[sql_queries_state, dataframes_state, plots_state],
@@ -276,7 +255,7 @@ def setup_drias_events(ui_elements: DriasUIElements, share_client=None, user_id=
         outputs=[ui_elements["details_accordion"]]
     ).then(
         ask_drias_query,
-        inputs=[ui_elements["drias_direct_question"], index_state],
+        inputs=[ui_elements["drias_direct_question"], index_state, user_id],
         outputs=[
             ui_elements["drias_sql_query"],
             ui_elements["drias_table"],
@@ -288,10 +267,6 @@ def setup_drias_events(ui_elements: DriasUIElements, share_client=None, user_id=
             table_names_list,
             ui_elements["result_text"],
         ],
-    ).then(
-        log_drias_interaction,
-        inputs=[ui_elements["drias_direct_question"], ui_elements["drias_sql_query"], ui_elements["drias_table"]],
-        outputs=[],
     ).then(
         show_results,
         inputs=[sql_queries_state, dataframes_state, plots_state],

front/utils.py

@@ -13,17 +13,6 @@ def make_pairs(lst:list)->list:
     return [(lst[i], lst[i + 1]) for i in range(0, len(lst), 2)]
 
 
-def serialize_docs(docs:list)->list:
-    new_docs = []
-    for doc in docs:
-        new_doc = {}
-        new_doc["page_content"] = doc.page_content
-        new_doc["metadata"] = doc.metadata
-        new_docs.append(new_doc)
-    return new_docs
-
-
-
 def parse_output_llm_with_sources(output:str)->str:
     # Split the content into a list of text and "[Doc X]" references
     content_parts = re.split(r'\[(Doc\s?\d+(?:,\s?Doc\s?\d+)*)\]', output)

requirements.txt

@@ -8,6 +8,7 @@ langgraph==0.2.70
 pinecone-client==4.1.0
 sentence-transformers==2.6.0
 huggingface-hub==0.25.2
+datasets==3.5.0
 pyalex==0.13
 networkx==3.2.1
 pyvis==0.3.2