DB storage for captions

Functions/caption_editor_functions.py

@@ -1,43 +1,29 @@
 import pandas as pd
-import json
-
-FILE_PATH = "Resources/captions.jsonl"
+from Functions.db_connection import default_app
 
 
 def get_captions_by_video_id(video_id):
-    with open(FILE_PATH) as file:
-        captions = pd.read_json(file, lines=True)
-
-    captions_edit = captions[captions['file'] == video_id]
-    captions_edit = captions_edit[['start_time', 'text', 'end_time']]
+    response = default_app.database().child("Captions").child(video_id).get().val()
+    captions = pd.DataFrame(response)
+    captions_edit = captions[['start_time', 'text', 'end_time']]
     captions_edit.columns = ["Start", "Text", "End"]
-    return captions_edit
-
+    return captions_edit, captions
 
-def save_dataframe(df, video_id, user):
-    cols = ["clean_text", "start_time", "user_id", "signer", "file", "end_time", "url", "text"]
-    other_captions_data = []
-    new_captions_data = []
 
-    with open(FILE_PATH) as f:
-        for line in f:
-            caption = json.loads(line)
-            if caption['file'] == video_id:
-                new_captions_data.append(caption)
-            else:
-                other_captions_data.append(caption)
-
-    other_captions = pd.DataFrame(data=other_captions_data, columns=cols)
-    new_captions = pd.DataFrame(data=new_captions_data, columns=cols)
+def save_dataframe(df, df_full, video_id, user):
     try:
-        new_captions['start_time'] = df['Start'].apply(lambda x: float(x))
-        new_captions['text'] = df['Text']
-        new_captions['end_time'] = df['End'].apply(lambda x: float(x))
-        new_captions['user_id'] = user
-
-        all_captions = pd.concat([other_captions, new_captions], ignore_index=True)
+        df_full["user_id"].loc[
+            df_full["start_time"] != df["Start"] or
+            df_full["end_time"] != df["End"] or
+            df_full["text"] != df["Text"]
+                                ] = user
+        df_full["start_time"] = df["Start"].apply(lambda x: float(x))
+        df_full["text"] = df["Text"]
+        df_full["end_time"] = df["End"].apply(lambda x: float(x))
+
+        df_json = df.to_json(orient="index")
+        default_app.database().child("Captions").child(video_id).set(df_json)
 
-        all_captions.to_json(FILE_PATH, orient='records', lines=True)
         return "Save successful!"
     except ValueError:
         return "Save failed: Incorrect input format"

app.py

@@ -13,8 +13,8 @@ def get_username(profile: gr.OAuthProfile):
     return profile
 
 
-def save(df, video_id):
-    return save_dataframe(df, video_id, user)
+def save(df, df_full, video_id):
+    return save_dataframe(df, df_full, video_id, user)
 
 
 def get_next_components():
@@ -28,12 +28,12 @@ def get_next_components():
     next_video_id = youtube_link_to_id(next_video_link)
 
     next_video = get_video_embed_by_id(next_video_id)
-    next_captions = get_captions_by_video_id(next_video_id)
+    next_captions, next_captions_full = get_captions_by_video_id(next_video_id)
 
-    return next_video, next_captions, next_video_id
+    return next_video, next_video_id, next_captions, next_captions_full
 
 
-(start_video, start_captions, start_video_id) = get_next_components()
+(start_video, start_video_id, start_captions, start_captions_full) = get_next_components()
 
 with gr.Blocks(css=css) as main_page:
     gr.Markdown("# Caption Editor")
@@ -42,6 +42,7 @@ with gr.Blocks(css=css) as main_page:
 
     current_user = gr.Textbox(visible=False, interactive=False)
     current_video_id = gr.Textbox(value=start_video_id, visible=False, interactive=False)
+    current_captions_full = gr.DataFrame(value=start_captions_full, visible=False, interactive=False)
 
     main_page.load(get_username, outputs=current_user)
 
@@ -64,9 +65,9 @@ with gr.Blocks(css=css) as main_page:
                     next_video_button = gr.Button("Next")
 
             next_video_button.click(fn=get_next_components,
-                                    outputs=[video_embed, caption_editor, current_video_id])
+                                    outputs=[video_embed, caption_editor, current_video_id, current_captions_full])
             save_button.click(fn=save,
-                              inputs=[caption_editor, current_video_id],
+                              inputs=[caption_editor, current_captions_full, current_video_id],
                               outputs=save_result)
 
 main_page.launch(share=True, ssr_mode=False)