KWOM_July

@@ -25,6 +25,15 @@ def extract_info(ocr_results):
 
 
 
 
 def predict(pdf_info, is_qrs=False, is_fsm=False):
 def predict(pdf_info, is_qrs=False, is_fsm=False):
+    pop_seceond_page_info = {}
+    if not is_fsm and not is_qrs and len(pdf_info) == 9:
+        pop_seceond_page_info = pdf_info.pop('1', {})
+        for pno in range(8):
+            if pno == 0:
+                pdf_info[str(pno)]['blocks'].extend(pop_seceond_page_info.get('blocks', []))
+            else:
+                pdf_info[str(pno)] = pdf_info.pop(str(pno+1))
+
     ocr_results = {}
     ocr_results = {}
     for pno in pdf_info:
     for pno in pdf_info:
         ocr_results[pno] = {}
         ocr_results[pno] = {}

@@ -13,6 +13,7 @@ class Finder:
         self.item = {"words": None,
         self.item = {"words": None,
                      "position": None,
                      "position": None,
                     }
                     }
+        self.cn_re = re.compile(u'[\u4e00-\u9fa5]')
 
 
     def gen_init_result(self, is_asp):
     def gen_init_result(self, is_asp):
         # 格式化算法输出
         # 格式化算法输出
@@ -187,6 +188,11 @@ class Finder:
                 vin['position'] = location
                 vin['position'] = location
         return vin
         return vin
 
 
+    def cn_char_filter(self, src_str):
+        cn_chars = re.findall(self.cn_re, src_str)
+        cn_str = ''.join(cn_chars)
+        return cn_str
+
     def get_loan_principal(self, page_num='0'):
     def get_loan_principal(self, page_num='0'):
         chinese_keywords = ['壹', '贰', '叁', '肆', '伍', '陆', '柒', '捌', '玖', '拾',
         chinese_keywords = ['壹', '贰', '叁', '肆', '伍', '陆', '柒', '捌', '玖', '拾',
                             '佰', '仟', '万', '亿', '元', '角', '分', '零', '整']
                             '佰', '仟', '万', '亿', '元', '角', '分', '零', '整']
@@ -201,7 +207,7 @@ class Finder:
             for line in block['lines']:
             for line in block['lines']:
                 for span in line['spans']:
                 for span in line['spans']:
                     bbox, text = span['bbox'], span['text']
                     bbox, text = span['bbox'], span['text']
-                    if fuzz.ratio(''.join(chinese_keywords), text) > 15:
+                    if fuzz.ratio(''.join(chinese_keywords), self.cn_char_filter(text)) >= 10:
                         text = text.split('：')[-1].strip()
                         text = text.split('：')[-1].strip()
                         upper['position'] = bbox
                         upper['position'] = bbox
                         upper['words'] = text
                         upper['words'] = text