use minio to store uploaded files; build dialog server; (#16)

* format code

* use minio to store uploaded files; build dialog server;

python/parser/docx_parser.py

@@ -3,6 +3,7 @@ import re
 import pandas as pd
 from collections import Counter
 from nlp import huqie
+from io import BytesIO
 
 
 class HuDocxParser:
@@ -97,7 +98,7 @@ class HuDocxParser:
         return ["\n".join(lines)]
 
     def __call__(self, fnm):
-        self.doc = Document(fnm)
+        self.doc = Document(fnm) if isinstance(fnm, str) else Document(BytesIO(fnm))
         secs = [(p.text, p.style.name) for p in self.doc.paragraphs]
         tbls = [self.__extract_table_content(tb) for tb in self.doc.tables]
         return secs, tbls

python/parser/excel_parser.py

@@ -1,10 +1,12 @@
 from openpyxl import load_workbook
 import sys
+from io import BytesIO
 
 
 class HuExcelParser:
     def __call__(self, fnm):
-        wb = load_workbook(fnm)
+        if isinstance(fnm, str):wb = load_workbook(fnm)
+        else: wb = load_workbook(BytesIO(fnm))
         res = []
         for sheetname in wb.sheetnames:
             ws = wb[sheetname]

python/parser/pdf_parser.py

@@ -1,4 +1,5 @@
 import xgboost as xgb
+from io import BytesIO
 import torch
 import re
 import pdfplumber
@@ -1525,7 +1526,7 @@ class HuParser:
         return "\n\n".join(res)
 
     def __call__(self, fnm, need_image=True, zoomin=3, return_html=False):
-        self.pdf = pdfplumber.open(fnm)
+        self.pdf = pdfplumber.open(fnm) if isinstance(fnm, str) else pdfplumber.open(BytesIO(fnm))
         self.lefted_chars = []
         self.mean_height = []
         self.mean_width = []