Upload /sim_case_sug_demo/g_h.py with huggingface_hub

sim_case_sug_demo/g_h.py

@@ -196,7 +196,7 @@ def ansi_to_html_dis(_f,file_path,_tranpose=True):
 
     df=pd.DataFrame(_dict)
     html_table_blue_light = build_table(df, 'blue_light')
-    #print(type(html_table_blue_light))
+    ##print(type(html_table_blue_light))
     injection="<meta charset=\"UTF-8\">"
     #"<td style = \"background-color: #D9E1F2;font-family: Century Gothic, sans-serif;font-size: medium;text-align: left;padding: 0px 20px 0px 0px;width: auto\">"
     html_table_blue_light=html_table_blue_light[:html_table_blue_light.find("<thead>")+7]+injection+html_table_blue_light[html_table_blue_light.find("<thead>")+7:]
@@ -226,7 +226,7 @@ def ansi_to_html(_f,file_path,_tranpose=True):
 
     df=pd.DataFrame(_dict)
     html_table_blue_light = build_table(df, 'blue_light')
-    #print(type(html_table_blue_light))
+    ##print(type(html_table_blue_light))
     injection="<meta charset=\"UTF-8\">"
     #"<td style = \"background-color: #D9E1F2;font-family: Century Gothic, sans-serif;font-size: medium;text-align: left;padding: 0px 20px 0px 0px;width: auto\">"
     html_table_blue_light=html_table_blue_light[:html_table_blue_light.find("<thead>")+7]+injection+html_table_blue_light[html_table_blue_light.find("<thead>")+7:]
@@ -294,7 +294,7 @@ def ansi_to_image(ansi_text, font_size=20, image_path="./test.png"):
         segments = line.split('\033')
         anchor_bg_color=(255,255,255)
         for segment in segments:
-            #print(segment)
+            ##print(segment)
             if segment and  segment[-1]=='m':
                 code= segment[:-1]
                 anchor_bg_color = ANSI_BG_COLORS.get(f'\033{code}m', anchor_bg_color)
@@ -345,7 +345,7 @@ def suggesting_dis(the_pool,target_name,case_dict):
     v_plst_2=[_embedder.encode(_e) for _e in plst_2]
     v_dlst_2=[_embedder.encode(_e) for _e in dlst_2]
 
-    print(clst_2)
+    #print(clst_2)
 
     rt_lst=[]
     for i in tqdm(the_pool):
@@ -355,13 +355,13 @@ def suggesting_dis(the_pool,target_name,case_dict):
         id_lst_1=[id_lst[sen_lst.index(_e)] for _e in lst_1]
         vec_lst_1=[vec_lst[sen_lst.index(_e)] for _e in lst_1]#[_embedder.encode(_e) for _e in lst_1]
         clst_1=[corpus_clust_label[_e] for _e in id_lst_1]#[clust_search(_cluster_core_dict,_e,0.68) for _e in vec_lst_1]
-        #print(clst_1)
+        ##print(clst_1)
         inset=sorted([_e for _e in set(clst_1)&set(clst_2) if _e!=-1])
         temp_ot={}
         if len(inset)>=max(1,inset_th):
             temp_ot["target"]=target_name
             temp_ot["inset"]=inset
-            #print(len(inset))
+            ##print(len(inset))
             _img=img_resize(vec2img(vec_lst_1,clst_1,vec_lst_2,clst_2,clust_th),cnn_len)
             cnn_pred=cnn_model.predict(np.array([_img])/255)
             
@@ -378,7 +378,7 @@ def suggesting_dis(the_pool,target_name,case_dict):
                     _con2.append([0]*emb_dim)
             _con1=np.array([_con1])
             _con2=np.array([_con2])
-            print(len(_con1),len(_con2),len(_con2[0]))
+            #print(len(_con1),len(_con2),len(_con2[0]))
             #_con1=list(np.array(vec_lst_1).reshape(len(lst_1)*emb_dim))+[0]*(emb_dim*(bilstm_len-len(lst_1))) if len(lst_1)<=bilstm_len else list(np.array(vec_lst_1).reshape(len(lst_1)*emb_dim))[:bilstm_len*emb_dim]
             #_con2=list(np.array(vec_lst_2).reshape(len(lst_2)*emb_dim))+[0]*(emb_dim*(bilstm_len-len(lst_2))) if len(lst_2)<=bilstm_len else list(np.array(vec_lst_2).reshape(len(lst_2)*emb_dim))[:bilstm_len*emb_dim]
             bilstm_pred=bilstm_model.predict([_con1,_con2])
@@ -386,14 +386,14 @@ def suggesting_dis(the_pool,target_name,case_dict):
 
             temp_ot["cnn_pred"]=float(cnn_pred[0][0])
             temp_ot["bilstm_pred"]=float(bilstm_pred[0][0])
-            #print(cnn_pred)
-            #print(bilstm_pred)
+            ##print(cnn_pred)
+            ##print(bilstm_pred)
             x_e=[[bilstm_pred[0][0],cnn_pred[0][0]]]
             if bilstm_pred[0][0]>=0.75:
                 ensemble_pred=logistic(x_r,y_r,x_e)
                 temp_ot["ensemble_pred"]=float(ensemble_pred[0])
 
-            #print(ensemble_pred)
+            ##print(ensemble_pred)
 
                 pre_lst_1=[[color_lst[inset.index(clst_1[_e]) % len(color_lst)],Fore.WHITE,lst_1[_e],Style.RESET_ALL] if clst_1[_e] in inset else [Style.RESET_ALL,lst_1[_e]] for _e in range(len(lst_1))]
                 pre_lst_2=[[color_lst[inset.index(clst_2[_e]) % len(color_lst)],Fore.WHITE,lst_2[_e],Style.RESET_ALL] if clst_2[_e] in inset else [Style.RESET_ALL,lst_2[_e]] for _e in range(len(lst_2))]
@@ -401,7 +401,7 @@ def suggesting_dis(the_pool,target_name,case_dict):
                 vlst_1=[[vec_lst_1[_e],pre_lst_1[_e][0]] for _e in range(len(pre_lst_1)) if len(pre_lst_1[_e])==4]
                 vlst_2=[[vec_lst_2[_e],pre_lst_2[_e][0]] for _e in range(len(pre_lst_2)) if len(pre_lst_2[_e])==4]
                 
-                #print(lst_1)
+                ##print(lst_1)
 
                 plst_1=replace_all("".join(corpus_pd_f[i.replace("_",",")][0]),key_lst,sp_key,1).split(sp_key)
                 
@@ -428,10 +428,10 @@ def suggesting_dis(the_pool,target_name,case_dict):
                 #if max_dp<max([len(plst_1),len(plst_2),len(dlst_1),len(dlst_2)]):
                 #   max_dp=max([len(plst_1),len(plst_2),len(dlst_1),len(dlst_2)])
                 
-                #print(plst_1)
-                #print(plst_2)
-                #print(dlst_1)
-                #print(dlst_2)
+                ##print(plst_1)
+                ##print(plst_2)
+                ##print(dlst_1)
+                ##print(dlst_2)
                 draw_lst_1=["".join(_e) for _e in pre_lst_1]
                 draw_lst_2=["".join(_e) for _e in pre_lst_2]
 
@@ -441,11 +441,11 @@ def suggesting_dis(the_pool,target_name,case_dict):
                 draw_lst_d2=["".join(_e) for _e in pre_lst_d2]
                 #replace_all(temp_c,key_lst,",",0)
 
-                #print(plst_1)
+                ##print(plst_1)
                 tp_str=""
 
-                #print("---------------------")
-                #print(Fore.BLUE+str(i)+Style.RESET_ALL)
+                ##print("---------------------")
+                ##print(Fore.BLUE+str(i)+Style.RESET_ALL)
                 temp_ot["case_id"]=i
                 temp_ot["plaintiff_case1"]=[{"background_color":ANSI_COLORS[_e[0]],"font_color":ANSI_COLORS[_e[1]],"content":_e[-2]} if len(_e)==4 else {"background_color":ANSI_COLORS[Back.WHITE],"font_color":ANSI_COLORS[Fore.BLACK],"content":_e[-1]} for _e in pre_lst_p1]
                 temp_ot["defendant_case1"]=[{"background_color":ANSI_COLORS[_e[0]],"font_color":ANSI_COLORS[_e[1]],"content":_e[-2]} if len(_e)==4 else {"background_color":ANSI_COLORS[Back.WHITE],"font_color":ANSI_COLORS[Fore.BLACK],"content":_e[-1]} for _e in pre_lst_d1]
@@ -489,7 +489,7 @@ def suggesting_dis(the_pool,target_name,case_dict):
                 
                 temp_ot["output"]=tp_str
                 rt_lst.append(temp_ot)
-                print(tp_str)
+                #print(tp_str)
     ot=sorted(rt_lst,key=lambda x:x["ensemble_pred"],reverse=True)
     ot_lst=[i["output"] for i in ot[:sug_th]]
 
@@ -514,7 +514,7 @@ def suggesting(the_pool,target_name,case_dict):
     v_plst_2=[_embedder.encode(_e) for _e in plst_2]
 
 
-    print(clst_2)
+    #print(clst_2)
 
     rt_lst=[]
     for i in tqdm(the_pool):
@@ -524,13 +524,13 @@ def suggesting(the_pool,target_name,case_dict):
         id_lst_1=[id_lst[sen_lst.index(_e)] for _e in lst_1]
         vec_lst_1=[vec_lst[sen_lst.index(_e)] for _e in lst_1]#[_embedder.encode(_e) for _e in lst_1]
         clst_1=[corpus_clust_label[_e] for _e in id_lst_1]#[clust_search(_cluster_core_dict,_e,0.68) for _e in vec_lst_1]
-        #print(clst_1)
+        ##print(clst_1)
         inset=sorted([_e for _e in set(clst_1)&set(clst_2) if _e!=-1])
         temp_ot={}
         if len(inset)>=max(1,inset_th):
             temp_ot["target"]=target_name
             temp_ot["inset"]=inset
-            #print(len(inset))
+            ##print(len(inset))
             _img=img_resize(vec2img(vec_lst_1,clst_1,vec_lst_2,clst_2,clust_th),cnn_len)
             cnn_pred=cnn_model.predict(np.array([_img])/255)
 
@@ -547,18 +547,18 @@ def suggesting(the_pool,target_name,case_dict):
                     _con2.append([0]*emb_dim)
             _con1=np.array([_con1])
             _con2=np.array([_con2])
-            print(len(_con1),len(_con2),len(_con2[0]))
+            #print(len(_con1),len(_con2),len(_con2[0]))
             #_con1=list(np.array(vec_lst_1).reshape(len(lst_1)*emb_dim))+[0]*(emb_dim*(bilstm_len-len(lst_1))) if len(lst_1)<=bilstm_len else list(np.array(vec_lst_1).reshape(len(lst_1)*emb_dim))[:bilstm_len*emb_dim]
             #_con2=list(np.array(vec_lst_2).reshape(len(lst_2)*emb_dim))+[0]*(emb_dim*(bilstm_len-len(lst_2))) if len(lst_2)<=bilstm_len else list(np.array(vec_lst_2).reshape(len(lst_2)*emb_dim))[:bilstm_len*emb_dim]
             bilstm_pred=bilstm_model.predict([_con1,_con2])
             temp_ot["cnn_pred"]=float(cnn_pred[0][0])
             temp_ot["bilstm_pred"]=float(bilstm_pred[0][0])
-            #print(cnn_pred)
-            #print(bilstm_pred)
+            ##print(cnn_pred)
+            ##print(bilstm_pred)
             x_e=[[bilstm_pred[0][0],cnn_pred[0][0]]]
             ensemble_pred=logistic(x_r,y_r,x_e)
             temp_ot["ensemble_pred"]=float(ensemble_pred[0])
-            #print(ensemble_pred)
+            ##print(ensemble_pred)
 
             pre_lst_1=[[color_lst[inset.index(clst_1[_e]) % len(color_lst)],Fore.WHITE,lst_1[_e],Style.RESET_ALL] if clst_1[_e] in inset else [Style.RESET_ALL,lst_1[_e]] for _e in range(len(lst_1))]
             pre_lst_2=[[color_lst[inset.index(clst_2[_e]) % len(color_lst)],Fore.WHITE,lst_2[_e],Style.RESET_ALL] if clst_2[_e] in inset else [Style.RESET_ALL,lst_2[_e]] for _e in range(len(lst_2))]
@@ -566,7 +566,7 @@ def suggesting(the_pool,target_name,case_dict):
             vlst_1=[[vec_lst_1[_e],pre_lst_1[_e][0]] for _e in range(len(pre_lst_1)) if len(pre_lst_1[_e])==4]
             vlst_2=[[vec_lst_2[_e],pre_lst_2[_e][0]] for _e in range(len(pre_lst_2)) if len(pre_lst_2[_e])==4]
             
-            #print(lst_1)
+            ##print(lst_1)
 
             plst_1=replace_all("".join(corpus_pd_f[i.replace("_",",")][0]),key_lst,sp_key,1).split(sp_key)
             
@@ -589,10 +589,10 @@ def suggesting(the_pool,target_name,case_dict):
             #if max_dp<max([len(plst_1),len(plst_2),len(dlst_1),len(dlst_2)]):
             #   max_dp=max([len(plst_1),len(plst_2),len(dlst_1),len(dlst_2)])
             
-            #print(plst_1)
-            #print(plst_2)
-            #print(dlst_1)
-            #print(dlst_2)
+            ##print(plst_1)
+            ##print(plst_2)
+            ##print(dlst_1)
+            ##print(dlst_2)
             draw_lst_1=["".join(_e) for _e in pre_lst_1]
             draw_lst_2=["".join(_e) for _e in pre_lst_2]
 
@@ -601,11 +601,11 @@ def suggesting(the_pool,target_name,case_dict):
 
             #replace_all(temp_c,key_lst,",",0)
 
-            #print(plst_1)
+            ##print(plst_1)
             tp_str=""
 
-            #print("---------------------")
-            #print(Fore.BLUE+str(i)+Style.RESET_ALL)
+            ##print("---------------------")
+            ##print(Fore.BLUE+str(i)+Style.RESET_ALL)
             temp_ot["case_id"]=i
 
             temp_ot["plaintiff_case1"]=[{"background_color":ANSI_COLORS[_e[0]],"font_color":ANSI_COLORS[_e[1]],"content":_e[-2]} if len(_e)==4 else {"background_color":ANSI_COLORS[Back.WHITE],"font_color":ANSI_COLORS[Fore.BLACK],"content":_e[-1]} for _e in pre_lst_p1]
@@ -642,7 +642,7 @@ def suggesting(the_pool,target_name,case_dict):
             
             temp_ot["output"]=tp_str
             rt_lst.append(temp_ot)
-            print(tp_str)
+            #print(tp_str)
     ot=sorted(rt_lst,key=lambda x:x["ensemble_pred"],reverse=True)
     ot_lst=[i["output"] for i in ot[:sug_th]]
 
@@ -794,11 +794,11 @@ import gradio as gr
 def case_sug_dis(file_name,plaintiff,defendant,p_point,d_point,dispute_list):
     global new_pd_f,new_point_f,corpus_dict
 
-    #print(file_name)
-    #print(point_f)
-    #print(list(pd_f.keys()).index(file_name))
+    ##print(file_name)
+    ##print(point_f)
+    ##print(list(pd_f.keys()).index(file_name))
     if file_name not in new_pd_f:
-        print("file not found")
+        #print("file not found")
         file_name="user_input"
     else:
         plaintiff=new_pd_f[file_name][0]
@@ -821,8 +821,8 @@ def case_sug_dis(file_name,plaintiff,defendant,p_point,d_point,dispute_list):
     dispute="\n".join(dispute_list)
     #ot=[Back.BLUE+dispute+Style.RESET_ALL]*10
     output_list=[]
-    print("-----")
-    print(len(ot_dict))
+    #print("-----")
+    #print(len(ot_dict))
     out_path="./out_of_range.html"
     for i in range(sug_th):
         if i<len(ot_dict):
@@ -836,11 +836,11 @@ def case_sug_dis(file_name,plaintiff,defendant,p_point,d_point,dispute_list):
 def case_sug(file_name,plaintiff,p_point):
     global new_pd_f,new_point_f,corpus_dict
 
-    print(file_name)
-    #print(point_f)
-    #print(list(pd_f.keys()).index(file_name))
+    #print(file_name)
+    ##print(point_f)
+    ##print(list(pd_f.keys()).index(file_name))
     if file_name not in new_pd_f:
-        print("file not found")
+        #print("file not found")
         file_name="user_input"
     else:
         plaintiff=new_pd_f[file_name][0]
@@ -852,15 +852,15 @@ def case_sug(file_name,plaintiff,p_point):
     p_point=p_point.split("。") if type(p_point)==type("111") else p_point
     _pool=[i for i in corpus_dict]
     _case_dict={"plaintiff":plaintiff,"p_point":p_point}
-    print(_case_dict,[type(_case_dict[_e]) for _e in _case_dict])
+    #print(_case_dict,[type(_case_dict[_e]) for _e in _case_dict])
     ot,ot_dict=suggesting(_pool,file_name,_case_dict)
 
 
 
     #ot=[Back.BLUE+dispute+Style.RESET_ALL]*10
     output_list=[]
-    print("-----")
-    print(len(ot_dict))
+    #print("-----")
+    #print(len(ot_dict))
     out_path="./out_of_range.html"
     for i in range(sug_th):
         if i<len(ot_dict):