added compress

app.py

@@ -3,6 +3,7 @@ import tempfile
 from typing import List
 import gradio as gr
 from PyPDF2 import PdfReader, PdfWriter
+import pikepdf
 
 # Load custom CSS
 with open(os.path.join(os.path.dirname(__file__), "fancy.css")) as f:
@@ -31,116 +32,97 @@ def parse_page_ranges(ranges: str, num_pages: int) -> List[int]:
 
 
 def split_pdf(file, page_ranges: str):
-    # Validate file size
-    file_size = os.path.getsize(file.name)
-    if file_size > MAX_SIZE_BYTES:
-        return None, f"File size exceeds 1 GB limit ({file_size / (1024*1024*1024):.2f} GB)."
-
+    if not file:
+        return None, "No file provided."
+    if os.path.getsize(file.name) > MAX_SIZE_BYTES:
+        return None, "File exceeds 1 GB limit."
     reader = PdfReader(file.name)
-    num_pages = len(reader.pages)
     try:
-        page_indices = parse_page_ranges(page_ranges, num_pages)
+        page_indices = parse_page_ranges(page_ranges, len(reader.pages))
     except Exception as e:
-        return None, f"Error parsing page ranges: {e}"
-
+        return None, str(e)
     if not page_indices:
-        return None, "No valid pages selected."
-
+        return None, "No valid pages."
     writer = PdfWriter()
-    for idx in page_indices:
-        writer.add_page(reader.pages[idx])
-
-    out_path = tempfile.NamedTemporaryFile(delete=False, suffix=".pdf").name
-    with open(out_path, "wb") as f_out:
-        writer.write(f_out)
-
-    return out_path, None
-
-
-def merge_pdfs(files):
-    # Merge multiple PDFs in upload order
-    if not files:
-        return None, "Please upload at least two PDF files to merge."
-
-    # Validate total size
-    total_size = sum(os.path.getsize(f.name) for f in files)
-    if total_size > MAX_SIZE_BYTES:
-        return None, f"Total file size exceeds 1 GB limit ({total_size / (1024*1024*1024):.2f} GB)."
-
+    for i in page_indices:
+        writer.add_page(reader.pages[i])
+    out = tempfile.NamedTemporaryFile(delete=False, suffix=".pdf").name
+    writer.write(out)
+    return out, None
+
+
+def merge_pdfs(files: List, order: List[int] = None):
+    if not files or len(files) < 2:
+        return None, "Upload at least two PDFs."
+    if sum(os.path.getsize(f.name) for f in files) > MAX_SIZE_BYTES:
+        return None, "Total exceeds 1 GB."
+    if order and len(order) == len(files):
+        files = [f for _, f in sorted(zip(order, files), key=lambda x: x[0])]
     writer = PdfWriter()
-    try:
-        for f in files:
-            reader = PdfReader(f.name)
-            for page in reader.pages:
-                writer.add_page(page)
-    except Exception as e:
-        return None, f"Error reading PDFs: {e}"
-
-    out_path = tempfile.NamedTemporaryFile(delete=False, suffix=".pdf").name
-    with open(out_path, "wb") as f_out:
-        writer.write(f_out)
+    for f in files:
+        for p in PdfReader(f.name).pages:
+            writer.add_page(p)
+    out = tempfile.NamedTemporaryFile(delete=False, suffix=".pdf").name
+    writer.write(out)
+    return out, None
 
-    return out_path, None
 
-# Build Gradio interface
-with gr.Blocks(css=custom_css) as demo:
-    gr.HTML("<h1 id='header'>PDF Splitter & Merger</h1>")
-    gr.Markdown(
-        "**Free & Quick Solution**: Process your PDFs in-memory instantly without storage.\n"
-        "Use the **Split** tab to extract specific pages or the **Merge** tab to combine multiple PDFs into one."
+def compress_pdf_stream(file):
+    reader = PdfReader(file.name)
+    writer = PdfWriter()
+    for p in reader.pages:
+        writer.add_page(p)
+    writer.compress_content_streams()
+    out = tempfile.NamedTemporaryFile(delete=False, suffix=".pdf").name
+    writer.write(out)
+    return out, None
+
+
+def compress_pdf_pike(file, quality: int = 75):
+    pdf = pikepdf.Pdf.open(file.name)
+    out = tempfile.NamedTemporaryFile(delete=False, suffix=".pdf").name
+    # optimize structure and recompress images with JPEG
+    pdf.save(
+        out,
+        optimize_version=True,
+        compression=pikepdf.Compression.jpeg,
+        jpeg_quality=quality,
+        linearize=True
     )
+    return out, None
 
+# Interface
+with gr.Blocks(css=custom_css) as demo:
+    gr.HTML("<h1>PDF Toolbox: Split, Merge & Compress</h1>")
+    gr.Markdown("**Free & Quick**: Manage PDFs in-memory. Privacy first.")
     with gr.Tabs():
-        # Split Tab
+        # Split
         with gr.TabItem("Split PDF"):
-            with gr.Row(elem_classes="input-row"):
-                pdf_input = gr.File(label="Select PDF to split", file_types=['.pdf'])
-                page_input = gr.Textbox(label="Page ranges", placeholder="e.g. 1-3,5,7-9")
-            with gr.Row(elem_classes="button-row"):
-                split_button = gr.Button("Split PDF", variant="primary")
-            output_split = gr.File(label="Download Split PDF")
-            error_split = gr.Textbox(label="Error Message", interactive=False, visible=False)
-
-            def run_split(file, ranges):
-                if file is None or not ranges:
-                    return None, "Please upload a PDF and specify page ranges.", True
-                out_path, error = split_pdf(file, ranges)
-                if error:
-                    return None, error, True
-                return out_path, "", False
-
-            split_button.click(
-                fn=run_split,
-                inputs=[pdf_input, page_input],
-                outputs=[output_split, error_split, error_split],
-                api_name="split_pdf"
-            )
-            error_split.change(lambda msg: msg != "", inputs=error_split, outputs=error_split)
+            inp, ranges = gr.File(label="PDF"), gr.Textbox(label="Ranges")
+            btn = gr.Button("Split")
+            out, err = gr.File(), gr.Textbox(visible=False)
+            btn.click(lambda f,r: (*split_pdf(f,r),), [inp,ranges],[out,err,err])
 
-        # Merge Tab
+        # Merge
         with gr.TabItem("Merge PDF"):
-            with gr.Row(elem_classes="input-row"):
-                merge_inputs = gr.Files(label="Select PDF files to merge", file_types=['.pdf'])
-            with gr.Row(elem_classes="button-row"):
-                merge_button = gr.Button("Merge PDFs", variant="primary")
-            output_merge = gr.File(label="Download Merged PDF")
-            error_merge = gr.Textbox(label="Error Message", interactive=False, visible=False)
-
-            def run_merge(files):
-                if not files or len(files) < 2:
-                    return None, "Please upload at least two PDF files.", True
-                out_path, error = merge_pdfs(files)
-                if error:
-                    return None, error, True
-                return out_path, "", False
-
-            merge_button.click(
-                fn=run_merge,
-                inputs=[merge_inputs],
-                outputs=[output_merge, error_merge, error_merge],
-                api_name="merge_pdfs"
-            )
-            error_merge.change(lambda msg: msg != "", inputs=error_merge, outputs=error_merge)
+            files = gr.Files(label="PDFs")
+            order_df = gr.Dataframe(headers=["Filename","Order"], interactive=True)
+            files.change(lambda fs: [[os.path.basename(f.name), i+1] for i,f in enumerate(fs)], inputs=files, outputs=order_df)
+            btn_m = gr.Button("Merge")
+            out_m, err_m = gr.File(), gr.Textbox(visible=False)
+            btn_m.click(lambda fs,df: merge_pdfs(fs,[int(r[1]) for r in df] if df else None), [files,order_df], [out_m,err_m,err_m])
+
+        # Compress
+        with gr.TabItem("Compress PDF"):
+            comp_file = gr.File(label="PDF to compress")
+            method = gr.Radio(choices=["Stream only","PikePDF"], value="PikePDF", label="Method")
+            quality = gr.Slider(10,100, value=75, step=5, label="JPEG Quality")
+            btn_c = gr.Button("Compress")
+            out_c, err_c = gr.File(), gr.Textbox(visible=False)
+            def do_compress(f,m,q):
+                if m=="PikePDF": return compress_pdf_pike(f,q)
+                return compress_pdf_stream(f)
+            btn_c.click(do_compress, [comp_file,method,quality], [out_c,err_c,err_c])
 
 if __name__ == "__main__":
-    demo.launch(server_name="0.0.0.0", server_port=int(os.environ.get("PORT", 7860)))
+    demo.launch(server_name="0.0.0.0", port=int(os.environ.get("PORT",7860)))