service_social.py

from entity import Docs, Cluster, Preprocess, SummaryInput
from fastapi import FastAPI
import time
import hashlib
import json
from fastapi.middleware.cors import CORSMiddleware
from function import topic_clustering_social as tc
# from function import topic_clustering_v2 as tc
from iclibs.ic_rabbit import ICRabbitMQ
from get_config import config_params

app = FastAPI()
app.add_middleware(
    CORSMiddleware,
    allow_origins=["*"],
    allow_credentials=True,
    allow_methods=["*"],
    allow_headers=["*"],
)


def get_hash_id(item: Docs):
    str_hash = ""
    for it in item.response["docs"]:
        str_hash += it["url"]
    str_hash += str(item.top_cluster)
    str_hash += str(item.top_sentence)
    str_hash += str(item.topn_summary)
    str_hash += str(item.top_doc)
    str_hash += str(item.threshold)
    if item.sorted_field.strip():
        str_hash += str(item.sorted_field)
    if item.delete_message:
        str_hash += str(item.delete_message)
    return hashlib.sha224(str_hash.encode("utf-8")).hexdigest()


try:
    with open("log_run/log.txt") as f:
        data_dict = json.load(f)
except Exception as ve:
    print(ve)
    data_dict = {}


@app.post("/newsanalysis/topic_clustering")
async def topic_clustering(item: Docs):
    docs = item.response["docs"]
    # threshold = item.threshold
    print("start ")
    print("len doc: ", len(docs))
    st_time = time.time()
    top_cluster = item.top_cluster
    top_sentence = item.top_sentence
    topn_summary = item.topn_summary
    sorted_field = item.sorted_field
    max_doc_per_cluster = item.max_doc_per_cluster
    hash_str = get_hash_id(item)
    # threshold = 0.1
    # item.threshold = threshold

    # with open("log/input_{0}.txt".format(st_time), "w+") as f:
    #     f.write(json.dumps({"docs": docs, "key": item.keyword}))

    print(hash_str)
    if len(docs) > 200:

        try:
            if hash_str in data_dict:
                path_res = data_dict[hash_str]["response_path"]
                with open(path_res) as ff:
                    results = json.load(ff)
                print("time analysis (cache): ", time.time() - st_time)
                return results
        except Exception as vee:
            print(vee)

    results = tc.topic_clustering(docs, item.threshold, top_cluster=top_cluster, top_sentence=top_sentence,
                                  topn_summary=topn_summary, sorted_field=sorted_field, max_doc_per_cluster=max_doc_per_cluster, delete_message=item.delete_message, is_check_spam=item.is_check_spam)

    path_res = "log/result_{0}.txt".format(hash_str)
    with open(path_res, "w+") as ff:
        ff.write(json.dumps(results))
    data_dict[hash_str] = {"time": st_time, "response_path": path_res}

    lst_rm = []
    for dt in data_dict:
        if time.time() - data_dict[dt]["time"] > 30 * 24 * 3600:
            lst_rm.append(dt)
    for dt in lst_rm:
        del data_dict[dt]
    with open("log_run/log.txt", "w+") as ff:
        ff.write(json.dumps(data_dict))
    print("time analysis: ", time.time() - st_time)
    return results