init

app.py

@@ -0,0 +1,149 @@
+import os
+import gradio as gr
+import logging
+import json
+import hashlib
+from pathlib import Path
+from fontTools.ttLib import TTFont, TTLibError
+from huggingface_hub import HfApi
+
+logging.basicConfig(level=logging.INFO)
+
+API = HfApi()
+TOKEN = os.environ.get("TOKEN", "your_token_here")  # fallback for debugging
+REPO_ID = "Felix92/docTR-resource-collection"
+
+def get_supported_chars(font_path: Path) -> list[str]:
+    try:
+        font = TTFont(font_path)
+        supported_chars = set()
+        for table in font["cmap"].tables:
+            supported_chars.update(table.cmap.keys())
+        chars = [chr(code_point) for code_point in sorted(supported_chars)]
+        return [char for char in chars if char.isprintable()]
+    except TTLibError as e:
+        logging.error(f"Error reading font file {font_path}: {e}")
+        return []
+    except Exception as e:
+        logging.error(f"Unexpected error reading font file {font_path}: {e}")
+        return []
+
+def get_sha256(file_path: Path) -> str:
+    hash_sha256 = hashlib.sha256()
+    with open(file_path, "rb") as f:
+        for chunk in iter(lambda: f.read(8192), b""):
+            hash_sha256.update(chunk)
+    return hash_sha256.hexdigest()
+
+def file_exists_on_hub(file_name: str, subfolder: str) -> bool:
+    files = API.list_repo_files(
+        repo_id=REPO_ID,
+        repo_type="dataset",
+        token=TOKEN,
+    )
+    return any(file.startswith(f"{subfolder}/{file_name}") for file in files)
+
+def _upload_hub(file_path: str, subfolder: str, sha_hash: str) -> None:
+    filename = f"{sha_hash}_{Path(file_path).name}"
+    repo_path = f"{subfolder}/{filename}"
+    API.upload_file(
+        path_or_fileobj=file_path,
+        path_in_repo=repo_path,
+        token=TOKEN,
+        repo_type="dataset",
+        repo_id=REPO_ID,
+    )
+    logging.info(f"Uploaded {repo_path}")
+
+def handle_uploads(font_upload, wordlist_upload, agree):
+    if not agree:
+        return gr.Markdown("You must agree to the terms and conditions before proceeding."), None, None, None
+
+    try:
+        if font_upload:
+            font_path = Path(font_upload)
+            font_sha = get_sha256(font_path)
+            if file_exists_on_hub(font_sha, "fonts"):
+                return gr.update(visible=False), gr.Markdown("<div style='text-align: center;'><h3>This font was already uploaded.</h3></div>"), gr.update(value=None), gr.update(value=None)
+
+            supported_chars = get_supported_chars(font_path)
+            if not supported_chars:
+                return gr.update(visible=False), gr.Markdown("<div style='text-align: center;'><h3>No supported characters found in the font file.</h3></div>"), gr.update(value=None), gr.update(value=None)
+            metadata = {
+                "font_name": font_path.stem,
+                "supported_characters": supported_chars,
+            }
+            json_path = font_path.with_suffix(".json")
+            with open(json_path, "w", encoding="utf-8") as f:
+                json.dump(metadata, f, ensure_ascii=False, indent=2)
+
+            json_sha = get_sha256(json_path)
+
+            _upload_hub(str(font_path), "fonts", font_sha)
+            _upload_hub(str(json_path), "fonts", json_sha)
+
+        if wordlist_upload:
+            wordlist_path = Path(wordlist_upload)
+            wordlist_sha = get_sha256(wordlist_path)
+            if file_exists_on_hub(wordlist_sha, "wordlists"):
+                return gr.update(visible=False), gr.Markdown("<div style='text-align: center;'><h3>This wordlist was already uploaded.</h3></div>"), gr.update(value=None), gr.update(value=None)
+
+            _upload_hub(str(wordlist_path), "wordlists", wordlist_sha)
+
+        return gr.update(visible=False), gr.Markdown("<div style='text-align: center;'><h3>Upload was successful! You can upload another item.</h3></div>"), gr.update(value=None), gr.update(value=None)
+
+    except Exception as e:
+        logging.exception("Upload failed")
+        return gr.update(visible=False), gr.Markdown(f"<div style='text-align: center;'><h3>An error occurred: {e}</h3></div>"), gr.update(value=None), gr.update(value=None)
+
+with gr.Blocks(fill_height=True) as demo:
+    agreement_markdown = gr.Markdown(
+        """
+        <div style="text-align: center;">
+        <h1>File Upload Agreement</h1>
+
+        <h3>This is a Hugging Face space for the docTR/OnnxTR community to collect wordlists and fonts for the following project/s:</h3>
+
+        <h3><a href="https://github.com/mindee/doctr">docTR</a></h3>
+
+        <h3><a href="https://github.com/felixdittrich92/OnnxTR">OnnxTR</a></h3>
+        </div>
+
+        <h3>The uploaded wordlists and fonts will be used to generate synthetic data.</h3>
+
+        <h3>All uploaded files can be found here: <a href="https://huggingface.co/datasets/Felix92/docTR-resource-collection">Hugging Face dataset</a></h3>
+
+        <br>
+        <br>
+
+        <h3>By uploading a wordlist or font, you explicitly agree to the following terms:</h3>
+
+        <h3>1. You affirm that you are the owner or have the necessary rights to upload and share the wordlist or font.</h3>
+
+        <h3>2. You agree that the uploaded wordlists / fonts will be made publicly available to everyone.</h3>
+
+        <h3>3. You agree that the uploaded wordlists / fonts can be used for any purpose, including commercial use, by any third party.</h3>
+        """
+    )
+    agree_button = gr.Button("I Agree to the Terms and Conditions")
+    agree_state = gr.State(value=False)
+
+    with gr.Column(visible=False) as upload_section:
+        success_message = gr.Markdown(visible=True)
+        font_upload = gr.File(label="Upload Font File [TTF | OTF]", file_types=[".ttf", ".otf"], type="filepath")
+        wordlist_upload = gr.File(label="Upload Wordlist [TXT]", file_types=[".txt"], type="filepath")
+        submit_button = gr.Button("Submit")
+
+    def toggle_agreement_visibility():
+        return gr.update(visible=False), gr.update(visible=False), True, gr.update(visible=True)
+
+    agree_button.click(fn=toggle_agreement_visibility, inputs=None, outputs=[agreement_markdown, agree_button, agree_state, upload_section])
+
+    submit_button.click(
+        fn=handle_uploads,
+        inputs=[font_upload, wordlist_upload, agree_state],
+        outputs=[agree_button, success_message, font_upload, wordlist_upload],
+    )
+
+if __name__ == "__main__":
+    demo.launch()

requirements.txt

@@ -0,0 +1,3 @@
+gradio>=5.0.0, <6.0.0
+fonttools>=4.57.0, <5.0.0
+huggingface-hub>=0.25.2, <1.0.0