Create app.py

app.py

@@ -0,0 +1,196 @@
+import gradio as gr
+import pandas as pd
+import io
+from datetime import date
+
+# Global variables
+sentences = []
+annotations = []
+categories = {}
+main_categories = []
+
+
+# ------------------------
+# Load categories.txt
+# ------------------------
+def load_categories(file):
+    global categories, main_categories
+    if file is None:
+        return "⚠️ No file uploaded", []
+
+    categories = {}
+    main_categories = []
+    current = None
+
+    with open(file.name, "r", encoding="utf-8") as f:
+        for line in f:
+            line = line.strip()
+            if not line:
+                continue
+            if line.endswith(":"):
+                current = line[:-1].strip()
+                main_categories.append(current)
+                categories[current] = []
+            elif current:
+                categories[current].append(line)
+
+    return "✅ Categories loaded", main_categories
+
+
+# ------------------------
+# Load sentences from TXT
+# ------------------------
+def load_txt(file):
+    global sentences, annotations
+    if file is None:
+        return "⚠️ No file uploaded", "", 0, 0, "", "", gr.update(visible=False)
+
+    with open(file.name, "r", encoding="utf-8") as f:
+        text = f.read()
+
+    sentences = [s.strip() for s in text.split(".") if s.strip()]
+    annotations = [["", "", ""] for _ in sentences]  # Cat A, Cat B, Comments
+
+    base_name = file.name.split("/")[-1].replace(".txt", "")
+    youtube_url = f"https://www.youtube.com/watch?v={base_name}"
+
+    return (
+        f"✅ {len(sentences)} sentences loaded",
+        sentences[0],
+        1,
+        len(sentences),
+        "",
+        youtube_url,
+        gr.update(visible=True),
+    )
+
+
+# ------------------------
+# Save annotation for current sentence
+# ------------------------
+def save_annotation(idx, catA, catB, comment):
+    global annotations
+    if 0 <= idx - 1 < len(annotations):
+        annotations[idx - 1] = [catA, catB, comment]
+    return "💾 Saved!"
+
+
+# ------------------------
+# Move to next/previous sentence
+# ------------------------
+def change_sentence(idx, direction, catA, catB, comment):
+    global sentences, annotations
+    save_annotation(idx, catA, catB, comment)
+
+    if direction == "next" and idx < len(sentences):
+        idx += 1
+    elif direction == "prev" and idx > 1:
+        idx -= 1
+
+    sentence = sentences[idx - 1]
+    catA, catB, comment = annotations[idx - 1]
+    return sentence, idx, catA, update_catB(catA), catB, comment
+
+
+# ------------------------
+# Update subcategories dynamically
+# ------------------------
+def update_catB(catA):
+    return gr.update(choices=categories.get(catA, []), value=None)
+
+
+# ------------------------
+# Export to Excel
+# ------------------------
+def export_excel(codeur, comments, youtube_link):
+    global sentences, annotations
+    rows = [
+        [s, ann[0], ann[1], ann[2]] for s, ann in zip(sentences, annotations)
+    ]
+    df = pd.DataFrame(rows, columns=["Sentence", "Category A", "Category B", "Comments"])
+
+    output = io.BytesIO()
+    with pd.ExcelWriter(output, engine="xlsxwriter") as writer:
+        df.to_excel(writer, index=False, sheet_name="Sentences")
+        meta = pd.DataFrame([[comments, youtube_link]], columns=["General Comments", "YouTube Link"])
+        meta.to_excel(writer, index=False, sheet_name="Metadata")
+    output.seek(0)
+
+    filename = f"Annotation_{codeur}_{date.today()}.xlsx"
+    return output, filename
+
+
+# ------------------------
+# Gradio UI
+# ------------------------
+with gr.Blocks() as demo:
+    gr.Markdown("# 🐱 My Little Cute Cat - Annotation Tool")
+
+    with gr.Tabs():
+        # -------- Instructions --------
+        with gr.Tab("📘 Instructions"):
+            gr.Markdown("""
+            ## How to Create Categories File
+            Write your categories in a text file like this:
+            ```
+            MainCategory1:
+            Subcat 1
+            Subcat 2
+
+            MainCategory2:
+            Subcat A
+            Subcat B
+            ```
+            Save as **categories.txt** and load it in the interface.
+            """)
+
+        # -------- Annotation --------
+        with gr.Tab("✍️ Annotation"):
+            status = gr.Label()
+
+            codeur = gr.Dropdown([f"Codeur {i}" for i in range(1, 11)], label="Select Codeur")
+            cat_file = gr.File(label="Load Categories File (.txt)", file_types=[".txt"])
+            file_input = gr.File(label="Upload TXT File (sentences)", file_types=[".txt"])
+
+            sentence_box = gr.Textbox(label="Sentence", interactive=False, lines=2)
+
+            with gr.Row():
+                catA = gr.Dropdown(label="Category A", choices=[])
+                catB = gr.Dropdown(label="Category B", choices=[])
+
+            comment = gr.Textbox(label="Comment for this sentence", lines=2)
+
+            with gr.Row():
+                prev_btn = gr.Button("⬅️ Previous")
+                save_btn = gr.Button("💾 Save")
+                next_btn = gr.Button("Next ➡️")
+
+            progress = gr.Label("")
+
+            general_comments = gr.Textbox(label="General Comments", lines=3)
+            youtube_link = gr.Textbox(label="YouTube Link")
+
+            export_btn = gr.Button("Export to Excel")
+            download_file = gr.File(label="Download Excel", visible=False)
+
+            # Events
+            cat_file.change(load_categories, inputs=cat_file, outputs=[status, catA])
+            file_input.change(load_txt, inputs=file_input,
+                              outputs=[status, sentence_box, progress, progress, comment, youtube_link, download_file])
+
+            catA.change(update_catB, inputs=catA, outputs=catB)
+
+            prev_btn.click(change_sentence, 
+                           inputs=[progress, gr.State("prev"), catA, catB, comment], 
+                           outputs=[sentence_box, progress, catA, catB, catB, comment])
+            next_btn.click(change_sentence, 
+                           inputs=[progress, gr.State("next"), catA, catB, comment], 
+                           outputs=[sentence_box, progress, catA, catB, catB, comment])
+
+            save_btn.click(save_annotation, inputs=[progress, catA, catB, comment], outputs=status)
+
+            export_btn.click(export_excel,
+                             inputs=[codeur, general_comments, youtube_link],
+                             outputs=[download_file, download_file])
+
+demo.launch()