wip

app.py

@@ -1,11 +1,11 @@
 import gradio as gr
 import os
+import setup
 
 temp_folder = 'tmp'
 
 
 def upload_file(filepath, source_lang, target_lang, translator_version):
-
     if translator_version == "v1":
         from src.salamandraTA7b_translator_external_aligner import SalamandraTA7bTranslatorExternalAligner
         from src.aligner import Aligner
@@ -60,4 +60,8 @@ with gr.Blocks() as demo:
                                                                                    [u, d, status_text])
         d.click(download_file, None, [u, d])
 if __name__ == "__main__":
+    # check if fastalign needs to be built, ie it's the first time this app is being run
+    if not os.path.exists("fast_align"):
+        setup.main()
+
     demo.launch(server_name="0.0.0.0", server_port=7860)

packages.txt

@@ -0,0 +1,4 @@
+libgomp1
+openjdk-21-jre-headless
+cmake
+build-essential

requirements.txt

@@ -11,4 +11,7 @@ torch~=2.8.0
 huggingface-hub~=0.36.0
 vllm~=0.11.0
 iso-639~=0.4.5
-accelerate~=1.11.0
+accelerate~=1.11.0
+gradio_client~=1.8.0
+gdown~=5.2.0
+GitPython~=3.1.45

setup.py

@@ -0,0 +1,97 @@
+import os
+import shutil
+import zipfile
+import subprocess
+import requests
+
+import sys
+
+from git import Repo
+import gdown
+
+
+def run(cmd, cwd=None):
+    """Run a system command with error checking."""
+    print(f"Running: {cmd}")
+    subprocess.run(cmd, check=True, cwd=cwd)
+
+
+def check_needs_building():
+    return not os.path.exists("fast_align") or not os.path.exists("atools") or not os.path.isdir(
+        "okapi-apps_gtk2-linux-x86_64_1.47.0")
+
+
+def main():
+    # -------------------------------------------------------------------
+    # 1. Clone and compile fast_align
+    # -------------------------------------------------------------------
+
+    print("Cloning fast_align repository...")
+    Repo.clone_from("https://github.com/clab/fast_align.git", "fast_align_src")
+
+    # Create build directory
+    build_dir = "fast_align_src/build"
+    os.makedirs(build_dir, exist_ok=True)
+
+    print("Running CMake...")
+    run(["cmake", "-DCMAKE_POLICY_VERSION_MINIMUM=3.5", ".."], cwd=build_dir)
+
+    print("Running make...")
+    run(["make"], cwd=build_dir)
+
+    # Copy binaries
+    shutil.copy("fast_align_src/build/fast_align", ".")
+    shutil.copy("fast_align_src/build/atools", ".")
+
+    # Cleanup
+    shutil.rmtree("fast_align_src")
+
+    # -------------------------------------------------------------------
+    # 2. Download fast_align config files
+    # -------------------------------------------------------------------
+
+    print("Downloading fast_align_config.zip using gdown...")
+    gdown.download(
+        "https://drive.google.com/uc?id=1OS-qbYLAgLJ2n4cpk9usNdTcNARjNaSr",
+        "fast_align_config.zip",
+        quiet=False
+    )
+
+    print("Unzipping config...")
+    with zipfile.ZipFile("fast_align_config.zip", "r") as z:
+        z.extractall(".")
+
+    os.remove("fast_align_config.zip")
+
+    # -------------------------------------------------------------------
+    # 3. Download Okapi Tikal
+    # -------------------------------------------------------------------
+
+    url = "https://okapiframework.org/binaries/main/1.47.0/okapi-apps_gtk2-linux-x86_64_1.47.0.zip"
+    outfile = "okapi-apps_gtk2-linux-x86_64_1.47.0.zip"
+
+    print("Downloading Okapi Tikal...")
+    with requests.get(url, stream=True) as r:
+        r.raise_for_status()
+        with open(outfile, "wb") as f:
+            for chunk in r.iter_content(chunk_size=8192):
+                f.write(chunk)
+
+    print("Unzipping Okapi Tikal...")
+    with zipfile.ZipFile(outfile, "r") as z:
+        z.extractall("okapi-apps_gtk2-linux-x86_64_1.47.0")
+
+    os.remove(outfile)
+
+    # -------------------------------------------------------------------
+    # 4. Install PspaCy model
+    # -------------------------------------------------------------------
+
+    print("Downloading spaCy model...")
+    run([sys.executable, "-m", "spacy", "download", "xx_ent_wiki_sm"])
+
+    print("\n All building tasks completed successfully!")
+
+
+if __name__ == '__main__':
+    main()