Upload 3 files

requirements.txt

@@ -0,0 +1,5 @@
+Gradio
+pandas
+PyMuPDF
+openpyxl
+lingtrain-aligner

src/app.py

@@ -0,0 +1,44 @@
+from gradio import Interface, File, Dropdown, Button, HTML
+import pandas as pd
+import os
+from utils import pdf_to_text, align_text
+
+def process_files(source_file, target_file, lang1, lang2):
+    if source_file is None or target_file is None:
+        return "Please upload both PDF files."
+
+    # Convert PDFs to text
+    text_content1 = pdf_to_text(source_file.name)
+    text_content2 = pdf_to_text(target_file.name)
+
+    # Align the texts
+    aligned_df = align_text(text_content1, text_content2, lang1, lang2)
+
+    # Convert DataFrame to HTML
+    aligned_html = aligned_df.to_html(index=False)
+
+    # Save DataFrame as Excel file
+    excel_path = "aligned_data.xlsx"
+    aligned_df.to_excel(excel_path, index=False)
+
+    return aligned_html, excel_path
+
+# Define the Gradio interface
+interface = Interface(
+    fn=process_files,
+    inputs=[
+        File(label="Upload Source PDF"),
+        File(label="Upload Target PDF"),
+        Dropdown(choices=["en", "es", "fr", "ch", "ar", "ru", "pt", "sw"], label="Select Language 1"),
+        Dropdown(choices=["en", "es", "fr", "ch", "ar", "ru", "pt", "sw"], label="Select Language 2"),
+    ],
+    outputs=[
+        HTML(label="Aligned DataFrame"),
+        Button(label="Download Aligned DataFrame")
+    ],
+    title="PDF Text Aligner",
+    description="Upload two PDF files and select languages to align the text."
+)
+
+if __name__ == "__main__":
+    interface.launch()

src/utils.py

@@ -0,0 +1,12 @@
+def pdf_to_text(pdf_path: str) -> str:
+    from PyPDF2 import PdfReader
+
+    text = ""
+    with open(pdf_path, "rb") as file:
+        reader = PdfReader(file)
+        for page in reader.pages:
+            text += page.extract_text() + "\n"
+    return text
+
+def save_to_excel(df, file_name: str):
+    df.to_excel(file_name, index=False)