Update app.py

app.py

@@ -1778,127 +1778,21 @@ def worker_mode_a(
         JOB_STATUS["running"] = False
 
 def worker_mode_b(
-    gguf_repo,
-    gguf_file,
-    imatrix_file,
-    profile,
-    target_repo,
+    bf16_path,
+    imatrix_path,
+    output_path,
+    quant_type
 ):
 
-    JOB_STATUS["running"] = True
-
-    try:
-
-        LOG_BUFFER.clear()
-
-        log("=" * 70)
-        log("🚀 MODE B START")
-        log("=" * 70)
-
-
-        if not gguf_repo or not gguf_file:
-
-            raise RuntimeError(
-                "❌ Missing GGUF repository or BF16 file"
-            )
-
-
-        # -----------------------------
-        # BF16 CACHE
-        # -----------------------------
-
-        # bf16 = get_or_download_bf16(
-        #     gguf_repo,
-        #     gguf_file
-        # )
-        bf16 = get_bucket_bf16( "Laguna-S-2.1-GGUF")
-
-   
-
-
-        log(
-            f"📦 BF16 SOURCE:\n{bf16}"
-        )
-
-
-
-        # -----------------------------
-        # IMATRIX
-        # -----------------------------
-
-        imatrix = None
-
-
-        if imatrix_file:
-
-            imatrix = download_imatrix(
-                gguf_repo,
-                imatrix_file
-            )
-
-            log(
-                f"📊 IMATRIX:\n{imatrix}"
-            )
-
-
-
-        # -----------------------------
-        # MODE B QUANT ONLY
-        # -----------------------------
-
-        gguf = quantize_mode_b(
-            bf16,
-            profile,
-            imatrix
-        )
-
-
-        size = (
-            os.path.getsize(gguf)
-            /
-            (1024 ** 3)
-        )
-
-
-        log(
-            f"📦 OUTPUT SIZE {size:.2f} GB"
-        )
-
-
-
-        # -----------------------------
-        # UPLOAD
-        # -----------------------------
-
-        upload_output(
-            gguf,
-            None,
-            target_repo
-        )
-
-
-        log("=" * 70)
-        log("🎉 MODE B COMPLETE")
-        log("=" * 70)
-
-
-
-    except Exception as e:
-
-        import traceback
-
-        log(
-            f"❌ MODE B ERROR: {e}"
-        )
-
-        log(
-            traceback.format_exc()
-        )
-
+    print("BF16:", bf16_path)
+    print("IMatrix:", imatrix_path)
 
-    finally:
-
-        JOB_STATUS["running"] = False
+    run_apex_quant(
+        input_path=bf16_path,
+        imatrix=imatrix_path,
+        output_path=output_path,
+        quant=quant_type
+    )
 def pipeline(
     mode,
     source_repo,
@@ -1970,12 +1864,13 @@ with gr.Blocks(title="APEX GGUF Factory") as demo:
 ## Mode A — HuggingFace Model
 `HF Model → F16 GGUF → APEX → Final GGUF`
 
-## Mode B — Existing BF16 GGUF
-`BF16 GGUF → Bucket Cache → APEX → Final GGUF`
+## Mode B — Bucket BF16 GGUF
+`Bucket BF16 GGUF → APEX → Final GGUF`
 
 ✅ Separate pipelines  
+✅ No downloads  
+✅ Direct bucket processing  
 ✅ Memory efficient  
-✅ Bucket cache support  
 ✅ APEX quantization  
 ✅ llama.cpp fallback
 """)
@@ -1988,7 +1883,7 @@ with gr.Blocks(title="APEX GGUF Factory") as demo:
     mode = gr.Radio(
         choices=[
             "Mode A - HF Model",
-            "Mode B - Existing BF16 GGUF"
+            "Mode B - Bucket BF16"
         ],
         value="Mode A - HF Model",
         label="Pipeline Mode"
@@ -2021,30 +1916,22 @@ with gr.Blocks(title="APEX GGUF Factory") as demo:
 
     with gr.Group(
         visible=False
-    ) as gguf_group:
+    ) as bucket_group:
 
         gr.Markdown(
-            "## 📦 Mode B: Existing BF16 GGUF"
-        )
-
-
-        gguf_repo = gr.Textbox(
-            label="GGUF Repository"
+            "## 📦 Mode B: Bucket BF16 GGUF"
         )
 
 
-        gguf_file = gr.Dropdown(
-            label="BF16 GGUF File",
-            choices=[],
-            interactive=True
+        bf16_path = gr.Textbox(
+            label="BF16 GGUF Bucket Path",
+            placeholder="/bucket/models/.../BF16"
         )
 
 
-        imatrix_file = gr.Dropdown(
-            label="IMATRIX (Optional)",
-            choices=[],
-            allow_custom_value=True,
-            interactive=True
+        imatrix_path = gr.Textbox(
+            label="I-Matrix Bucket Path (Optional)",
+            placeholder="/bucket/models/.../imatrix"
         )
 
 
@@ -2067,9 +1954,9 @@ with gr.Blocks(title="APEX GGUF Factory") as demo:
     )
 
 
-    target_repo = gr.Textbox(
-        label="Output HuggingFace Repository",
-        value="rahul7star/gemma-gguf"
+    output_path = gr.Textbox(
+        label="Output Bucket Path",
+        value="/bucket/output"
     )
 
 
@@ -2115,13 +2002,10 @@ with gr.Blocks(title="APEX GGUF Factory") as demo:
                 )
             )
 
-
         return (
-
             gr.update(
                 visible=False
             ),
-
             gr.update(
                 visible=True
             )
@@ -2133,22 +2017,7 @@ with gr.Blocks(title="APEX GGUF Factory") as demo:
         inputs=mode,
         outputs=[
             hf_group,
-            gguf_group
-        ]
-    )
-
-
-
-    # ==================================================
-    # MODE B REPO SCAN
-    # ==================================================
-
-    gguf_repo.change(
-        load_repo_files,
-        inputs=gguf_repo,
-        outputs=[
-            gguf_file,
-            imatrix_file
+            bucket_group
         ]
     )
 
@@ -2163,11 +2032,10 @@ with gr.Blocks(title="APEX GGUF Factory") as demo:
         inputs=[
             mode,
             source_repo,
-            gguf_repo,
-            gguf_file,
-            imatrix_file,
+            bf16_path,
+            imatrix_path,
             profile,
-            target_repo
+            output_path
         ],
         outputs=status
     )
@@ -2188,8 +2056,6 @@ with gr.Blocks(title="APEX GGUF Factory") as demo:
         outputs=logs
     )
 
-
-
 # ==================================================
 # LAUNCH
 # ==================================================