| import fitz |
| import os |
| import streamlit as st |
| from io import BytesIO |
|
|
| PDF_STORAGE_DIR = 'pdf_storage' |
|
|
| def parse_pdf(pdf_file_name, page_numbers, search_strings, rect_coords): |
| doc = fitz.open(os.path.join(PDF_STORAGE_DIR, pdf_file_name)) |
| pages = [doc.load_page(i) for i in page_numbers] |
| output = "" |
|
|
| output += "### Extracted Text:\n" |
| for page in pages: |
| output += "### Page %d\n" % page.number |
| output += page.get_text() + "\n\n" |
|
|
| if search_strings: |
| output += "\n### Search Results:\n" |
| for search_string in search_strings: |
| for page in pages: |
| rect = page.search_for(search_string) |
| if rect: |
| output += f"#### {search_string} - \n\n" |
| for r in rect: |
| r = list(map(lambda x: round(x,3), r)) |
| output += f"##### {r} \n\n" |
| else: |
| output += f"{search_string} not found\n" |
| output += "\n" |
|
|
| if rect_coords: |
| x1, y1, x2, y2 = map(float, rect_coords.split(',')) |
| rect = fitz.Rect(x1, y1, x2, y2) |
| output += "\n### Rectangle Text:\n" |
| for page in pages: |
| output += page.get_text("text", clip=rect) + "\n\n" |
| return output |
|
|
| st.write("### Extract Text from PDF") |
| |
| with st.container(): |
| pdf_files = os.listdir(PDF_STORAGE_DIR) |
| |
| pdf_file = st.selectbox("Select PDF file to delete", pdf_files) |
|
|
| st.write("### Enter Page Numbers:") |
| page_numbers = st.text_input("Enter page numbers (comma-separated)") |
|
|
| st.write("### Enter Search Strings:") |
| search_strings = st.text_input("Enter search strings (comma-separated)") |
|
|
| |
| st.write("### Enter Rectangle Coordinates:") |
| coord_method = st.selectbox("Select input method", ["Enter all at once", "Enter individually"]) |
|
|
| rect_coords = "" |
|
|
| if coord_method == "Enter all at once": |
| rect_coords = st.text_input("Enter rectangle coordinates (x1,y1,x2,y2)", placeholder="e.g., 10,20,30,40") |
| else: |
| col1, col2 = st.columns(2) |
| |
| with col1: |
| x1 = st.number_input("X1 (top-left x-coordinate)", min_value=0, value=0) |
| y1 = st.number_input("Y1 (top-left y-coordinate)", min_value=0, value=0) |
| |
| with col2: |
| x2 = st.number_input("X2 (bottom-right x-coordinate)", min_value=0, value=500) |
| y2 = st.number_input("Y2 (bottom-right y-coordinate)", min_value=0, value=500) |
| |
| rect_coords = f"{x1},{y1},{x2},{y2}" |
| |
| if st.button("Extract Text"): |
| if not page_numbers: |
| page_numbers = [0] |
| else: |
| page_numbers = [int(i) for i in page_numbers.split(',')] |
| |
| if not search_strings: |
| search_strings = [] |
| else: |
| search_strings = [s.strip() for s in search_strings.split(',')] |
| |
| output = parse_pdf(pdf_file, page_numbers, search_strings, rect_coords) |
| st.write(output) |
