fastapi_app.py

from fastapi import FastAPI, HTTPException
from semantic_search import SemanticSearch
from transaction_maps_search import TransactionMapsSearch
from pydantic import BaseModel
import os
import datetime
import json
import traceback
from llm.common import LlmParams, LlmPredictParams
from llm.deepinfra_api import DeepInfraApi

# Check if logs are enabled
ENABLE_LOGS = os.getenv("ENABLE_LOGS", "0") == "1"

# Set the path for log files
LOGS_BASE_PATH = os.getenv("LOGS_BASE_PATH", "logs")

# Create logs directory if it doesn't exist
if ENABLE_LOGS and not os.path.exists(LOGS_BASE_PATH):
    os.makedirs(LOGS_BASE_PATH)

LLM_API_URL = os.getenv("LLM_API_URL", "")
LLM_API_KEY = os.getenv("LLM_API_KEY", "")
LLM_USE_DEEPINFRA = os.getenv("LLM_USE_DEEPINFRA", "") == "1"

class Query(BaseModel):
    query: str = ''
    top: int = 10
    use_qe: bool = False
    use_olympic: bool = False
    find_transaction_maps_by_question: bool = False
    find_transaction_maps_by_operation: bool = False
    request_id: str = ''
    categories: dict = {'НКРФ': False,
                        'ГКРФ': False,
                        'ТКРФ': False,
                        'Федеральный закон': False,
                        'Письмо Минфина': False,
                        'Письмо ФНС': False,
                        'Приказ ФНС': False,
                        'Постановление Правительства': False,
                        'Судебный документ': False,
                        'ВНД': False,
                        'Бухгалтерский документ': False}
    llm_params: LlmParams = None

search = SemanticSearch()
transaction_maps_search = TransactionMapsSearch()

app = FastAPI(
    title="multistep-semantic-search-app",
    description="multistep-semantic-search-app",
    version="0.1.0",
)


def log_query_result(query, top, request_id, result):
    if not ENABLE_LOGS:
        return

    timestamp = datetime.datetime.now().strftime("%Y%m%d%H%M%S")
    log_file_path = os.path.join(LOGS_BASE_PATH, f"{timestamp}.json")

    log_data = {
        "timestamp": timestamp,
        "query": query,
        "top": top,
        "request_id": request_id,
        "result": result
    }

    with open(log_file_path, 'w', encoding='utf-8') as log_file:
        json.dump(log_data, log_file, indent=2, ensure_ascii=False)


@app.post('/search')
async def search_route(query: Query) -> dict:

    default_llm_params = LlmParams(url=LLM_API_URL,api_key=LLM_API_KEY, model="meta-llama/Llama-3.3-70B-Instruct", predict_params=LlmPredictParams(temperature=0.15, top_p=0.95, min_p=0.05, seed=42, repetition_penalty=1.2, presence_penalty=1.1, max_tokens=6000))

    try:
        question = getattr(query, "query", None)
        if not question:
            raise ValueError("Query parameter 'query' is required and cannot be empty.")

        top = getattr(query, "top", 15)
        use_qe = getattr(query, "use_qe", False)
        request_id = getattr(query, "request_id", None)
        categories = getattr(query, "categories", None)
        use_olympic = getattr(query, "use_olympic", False)
        find_transaction_maps_by_question = getattr(query, "find_transaction_maps_by_question", False)
        find_transaction_maps_by_operation = getattr(query, "find_transaction_maps_by_operation", False)

        request_llm_params = getattr(query, "llm_params", None)
        
        print(request_llm_params)
        
        llm_params = default_llm_params#getattr(query, "llm_params", default_llm_params)
            
        if LLM_USE_DEEPINFRA:
            print(llm_params.model)
            llm_api = DeepInfraApi(llm_params)
                                   
        
        if find_transaction_maps_by_question or find_transaction_maps_by_operation:
            transaction_maps_results, answer = await transaction_maps_search.search_transaction_map(
                query=question,
                find_transaction_maps_by_question=find_transaction_maps_by_question,
                k_neighbours=top,
                llm_api=llm_api)

            response = {'transaction_maps_results': transaction_maps_results}

        else:
            modified_query, titles, concat_docs, \
            relevant_consultations, predicted_explanation, \
            llm_responses = await search.search(question, use_qe, use_olympic, categories, llm_params)

            results = [{'title': str(item1), 'text_for_llm': str(item2)} for item1, item2 in
                        zip(titles, concat_docs)]

            consultations = [{'title': key, 'text': value} for key, value in relevant_consultations.items()]
            explanations = [{'title': key, 'text': value} for key, value in predicted_explanation.items()]

            response = {'query': modified_query, 'results': results,
                        'consultations': consultations, 'explanations': explanations, 'llm_responses': llm_responses}

        log_query_result(question, top, request_id, response)

        return response
    except ValueError as ve:
        traceback.print_exception(type(ve), ve, ve.__traceback__)
        raise HTTPException(status_code=400, detail=str(ve))
    except Exception as e:
        traceback.print_exception(type(e), e, e.__traceback__)
        raise HTTPException(status_code=500, detail=f"Unexpected error: {str(e)}")


@app.get('/health')
def health():
    return {"status": "ok"}


@app.get('/read_logs')
def read_logs():
    logs = []
    for log_file in os.listdir(LOGS_BASE_PATH):
        if log_file.endswith(".json"):
            with open(os.path.join(LOGS_BASE_PATH, log_file), 'r', encoding='utf-8') as file:
                log_data = json.load(file)
                logs.append(log_data)
    return logs


@app.get('/analyze_logs')
def analyze_logs():
    logs_by_query_top = {}
    for log_file in os.listdir(LOGS_BASE_PATH):
        if log_file.endswith(".json"):
            with open(os.path.join(LOGS_BASE_PATH, log_file), 'r', encoding='utf-8') as file:
                log_data = json.load(file)
                query = log_data.get("query", "")
                top = log_data.get("top", "")
                request_id = log_data.get("request_id", "")
                # Group logs by query and top
                key = f"{query}_{top}"
                if key not in logs_by_query_top:
                    logs_by_query_top[key] = []
                logs_by_query_top[key].append(log_data)

    # Analyze logs and filter out logs with different results for the same query and top
    invalid_logs = []
    for key, logs in logs_by_query_top.items():
        if len(set(json.dumps(log['result']) for log in logs)) > 1:
            invalid_logs.extend(logs)

    return invalid_logs