InitialMarkups2 / findspecsv1.py
Marthee's picture
Update findspecsv1.py
47356e3 verified
# -*- coding: utf-8 -*-
"""FindSpecsTrial(Retrieving+boundingBoxes).ipynb
Automatically generated by Colab.
Original file is located at
https://colab.research.google.com/drive/1mFuB1gtGuVh3NlOnNTzOFnDVuWSwn18q
"""
import fitz # PyMuPDF
from io import BytesIO
import re
import requests
import pandas as pd
from collections import Counter
import fitz # PyMuPDF
import re
import urllib.parse
import pandas as pd
import math
import random
# import tempfile
# from fpdf import FPDF
import json
from datetime import datetime
baselink='https://marthee-nbslink.hf.space/view-pdf?'
def get_repeated_texts(pdf_document, threshold=0.85):
"""
Identify text that appears on most pages, with font size and color.
:param pdf_document: The opened PDF document.
:param threshold: The percentage of pages a text must appear on to be considered "repeated".
:return: A list of dictionaries with text, font size, and color.
"""
text_counts = Counter()
text_metadata = defaultdict(list)
total_pages = pdf_document.page_count
for page_num in range(total_pages):
page = pdf_document.load_page(page_num)
blocks = page.get_text("dict")["blocks"]
seen_texts = set() # To avoid counting the same text twice per page
for block in blocks:
if "lines" not in block:
continue
for line in block["lines"]:
for span in line["spans"]:
text = span["text"].strip()
if not text:
continue
if text not in seen_texts:
seen_texts.add(text)
text_counts[text] += 1
text_metadata[text].append({
"font_size": span.get("size"),
"color": span.get("color")
})
# Find texts that appear in at least `threshold * total_pages` pages
min_occurrence = max(2, int(threshold * total_pages))
repeated_texts_info = []
for text, count in text_counts.items():
if count >= min_occurrence:
sizes = [meta["font_size"] for meta in text_metadata[text]]
colors = [meta["color"] for meta in text_metadata[text]]
# Get the most common size and color used for this text
most_common_size = max(set(sizes), key=sizes.count)
most_common_color = max(set(colors), key=colors.count)
repeated_texts_info.append({
"text": text,
"font_size": most_common_size,
"color": most_common_color
})
return repeated_texts_info
def get_regular_font_size_and_color(doc):
font_sizes = []
colors = []
fonts = []
# Loop through all pages
for page_num in range(len(doc)):
page = doc.load_page(page_num)
for span in page.get_text("dict")["blocks"]:
if "lines" in span:
for line in span["lines"]:
for span in line["spans"]:
font_sizes.append(span['size'])
colors.append(span['color'])
fonts.append(span['font'])
# Get the most common font size, color, and font
most_common_font_size = Counter(font_sizes).most_common(1)[0][0] if font_sizes else None
most_common_color = Counter(colors).most_common(1)[0][0] if colors else None
most_common_font = Counter(fonts).most_common(1)[0][0] if fonts else None
return most_common_font_size, most_common_color, most_common_font
import re
from collections import defaultdict
import fitz # PyMuPDF
import requests
from io import BytesIO
def normalize_text(text):
return re.sub(r'\s+', ' ', text.strip().lower())
def get_spaced_text_from_spans(spans):
return normalize_text(" ".join(span["text"].strip() for span in spans))
def is_header(span, most_common_font_size, most_common_color, most_common_font):
fontname = span.get("font", "").lower()
is_italic = "italic" in fontname or "oblique" in fontname
is_bold = "bold" in fontname or span.get("bold", False)
return (
not is_italic and (
span["size"] > most_common_font_size or
# span["color"] != most_common_color or
span["font"].lower() != most_common_font.lower() or
is_bold
)
)
def merge_consecutive_words(headers):
result = []
i = 0
while i < len(headers):
if i + 1 < len(headers) and headers[i] + ' ' + headers[i + 1] in headers:
result.append(headers[i] + ' ' + headers[i + 1])
i += 2
else:
result.append(headers[i])
i += 1
return result
def extract_headers(doc, toc_pages, most_common_font_size, most_common_color, most_common_font, top_margin, bottom_margin):
print("Font baseline:", most_common_font_size, most_common_color, most_common_font)
grouped_headers_by_y = defaultdict(list)
for pageNum in range(len(doc)):
if pageNum in toc_pages:
continue
page = doc.load_page(pageNum)
page_height = page.rect.height
text_instances = page.get_text("dict")
for block in text_instances['blocks']:
if block['type'] != 0:
continue
for line in block['lines']:
for span in line['spans']:
span_y = round(span['bbox'][1])
span_text = normalize_text(span.get('text', ''))
span_y0 = span['bbox'][1] # Top Y of this span
span_y1 = span['bbox'][3] # Bottom Y of this span
if span_y0 < top_margin or span_y1 > (page_height - bottom_margin):
continue
if not span_text:
continue
if span_text.startswith('http://www') or span_text.startswith('www'):
continue
if any((
'page' in span_text,
not re.search(r'[a-z0-9]', span_text),
'end of section' in span_text,
re.search(r'page\s+\d+\s+of\s+\d+', span_text),
re.search(r'\b(?:\d{1,2}[/-])?\d{1,2}[/-]\d{2,4}\b', span_text),
re.search(r'\b(?:jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)', span_text),
'specification:' in span_text
)):
continue
span_text = re.sub(r'[.\-]{4,}.*$', '', span_text).strip()
span_text = normalize_text(span_text)
if is_header(span, most_common_font_size, most_common_color, most_common_font):
grouped_headers_by_y[(pageNum, span_y)].append({
"text": span_text,
"size": span["size"],
"pageNum": pageNum
})
headers = []
for (pageNum, y), spans in sorted(grouped_headers_by_y.items()):
combined_text = " ".join(span['text'] for span in spans)
first_span = spans[0]
headers.append([combined_text, first_span['size'], first_span['pageNum'], y]) # <--- ADDED 'y'
# Analyze font sizes
font_sizes = [size for _, size, _, _ in headers] # <--- UNPACK 4 items now
font_size_counts = Counter(font_sizes)
top_3_font_sizes = sorted(font_size_counts.keys(), reverse=True)[:3]
return headers, top_3_font_sizes
class ColorManager:
def __init__(self, palette, min_distance=100):
self.palette = palette.copy()
self.used_colors = palette.copy()
self.idx = 0
self.min_distance = min_distance
def color_distance(self, c1, c2):
return math.sqrt(sum((a - b) ** 2 for a, b in zip(c1, c2)))
def generate_new_color(self):
max_attempts = 1000
for _ in range(max_attempts):
new_color = (random.randint(0, 255), random.randint(0, 255), random.randint(0, 255))
if all(self.color_distance(new_color, existing) > self.min_distance for existing in self.used_colors):
self.used_colors.append(new_color)
return new_color
raise ValueError("Couldn't find a distinct color after many attempts.")
def get_next_color(self):
if self.idx < len(self.palette):
color = self.palette[self.idx]
else:
color = self.generate_new_color()
self.idx += 1
return color
# Your original color palette
color_palette = [
(255, 0, 0), (0, 0, 255), (0, 255, 255), (0, 64, 0), (255, 204, 0),
(255, 128, 64), (255, 0, 128), (255, 128, 192), (128, 128, 255),
(128, 64, 0), (0, 255, 0), (0, 200, 0), (255, 128, 255), (128, 0, 255),
(0, 128, 192), (128, 0, 128), (128, 0, 0), (0, 128, 255), (149, 1, 70),
(255, 182, 128), (222, 48, 71), (240, 0, 112), (255, 0, 255),
(192, 46, 65), (0, 0, 128), (0, 128, 64), (255, 255, 0), (128, 0, 80),
(255, 255, 128), (90, 255, 140), (255, 200, 20), (91, 16, 51),
(90, 105, 138), (114, 10, 138), (36, 82, 78), (225, 105, 190),
(108, 150, 170), (11, 35, 75), (42, 176, 170), (255, 176, 170),
(209, 151, 15), (81, 27, 85), (226, 106, 122), (67, 119, 149),
(159, 179, 140), (159, 179, 30), (255, 85, 198), (255, 27, 85),
(188, 158, 8), (140, 188, 120), (59, 61, 52), (65, 81, 21),
(212, 255, 174), (15, 164, 90), (41, 217, 245), (213, 23, 182),
(11, 85, 169), (78, 153, 239), (0, 66, 141), (64, 98, 232),
(140, 112, 255), (57, 33, 154), (194, 117, 252), (116, 92, 135),
(74, 43, 98), (188, 13, 123), (129, 58, 91), (255, 128, 100),
(171, 122, 145), (255, 98, 98), (222, 48, 77)
]
# Create ONE color manager and re-use it
color_manager = ColorManager(color_palette)
def highlight_boxes(doc, highlights,color):
for page_num, bbox in highlights.items():
page = doc.load_page(page_num)
rect = fitz.Rect(bbox)
annot = page.add_rect_annot(rect)
rgb_color = tuple(c / 255 for c in color) # Normalize
annot.set_colors(stroke=rgb_color, fill=rgb_color)
annot.set_opacity(0.3)
annot.update()
def find_full_line_in_toc(doc, toc_pages, substring):
substring = normalize_text(substring) # Normalize for matching
best_match = None
for page_num in toc_pages:
page = doc.load_page(page_num)
blocks = page.get_text("dict")["blocks"]
for block in blocks:
for line in block.get("lines", []):
line_text = get_spaced_text_from_spans(line.get("spans", [])).strip()
normalized_line = normalize_text(line_text)
if substring in normalized_line:
# Remove dots and anything after
line_text = re.split(r'\.{2,}', line_text)[0].strip()
best_match = line_text
return best_match # stop at first match
return None
def extract_section_under_header(pdf_path, target_header_LIST):
top_margin=70
bottom_margin=50
df = pd.DataFrame(columns=["NBSLink","Subject","Page","Author","Creation Date","Layer",'Code', 'head above 1', "head above 2"])
dictionaryNBS={}
data_list_JSON = []
if pdf_path and ('http' in pdf_path or 'dropbox' in pdf_path):
pdf_path = pdf_path.replace('dl=0', 'dl=1')
response = requests.get(pdf_path)
pdf_content = BytesIO(response.content)
if not pdf_content:
raise ValueError("No valid PDF content found.")
doc = fitz.open(stream=pdf_content, filetype="pdf")
most_common_font_size, most_common_color, most_common_font =get_regular_font_size_and_color(doc)
def get_toc_page_numbers(doc, max_pages_to_check=15):
toc_pages = []
for page_num in range(min(len(doc), max_pages_to_check)):
page = doc.load_page(page_num)
blocks = page.get_text("dict")["blocks"]
dot_line_count = 0
lines_with_numbers_at_end = 0
for block in blocks:
for line in block.get("lines", []):
line_text = get_spaced_text_from_spans(line["spans"]).strip()
if re.search(r'\.{3,}', line_text):
dot_line_count += 1
# if re.search(r'\s\d{1,3}$', line_text):
# lines_with_numbers_at_end += 1
if dot_line_count >= 3 :#or lines_with_numbers_at_end >= 4:
toc_pages.append(page_num)
if bool(toc_pages):
return list(range(0, toc_pages[-1] + 1))
return toc_pages
toc_pages = get_toc_page_numbers(doc)
headers,top_3_font_sizes=extract_headers(doc,toc_pages,most_common_font_size, most_common_color, most_common_font,top_margin,bottom_margin)
if top_3_font_sizes:
mainHeaderFontSize, subHeaderFontSize, subsubheaderFontSize = top_3_font_sizes
print("Detected headers:", headers)
headers_set = set()
headers_dict = {}
for h in headers:
norm_text = normalize_text(h[0]) # h[0] is the text
headers_set.add(norm_text)
headers_dict[norm_text] = (h[0], h[1], h[2]) # (text, size, pageNum)
results = {}
print("📌 Has TOC:", bool(toc_pages), " | Pages to skip:", toc_pages)
matched_header_line = None # <-- Will store the line that acts as header
for heading_to_search in target_header_LIST:
print('headertosearch',heading_to_search)
matched_header_line = None
done=False
target_header = normalize_text(heading_to_search)
if target_header not in headers_set:
print(f"Header '{target_header}' not found. Searching for best match...")
heading_words = set(target_header.split())
best_match_score = 0
for page_num in range(len(doc)):
page = doc.load_page(page_num)
blocks = page.get_text("dict")["blocks"]
for block in blocks:
for line in block.get("lines", []):
line_text = " ".join(span["text"].strip() for span in line.get("spans", []))
if not line_text:
continue
line_words = set(re.findall(r'\w+', line_text.lower()))
match_count = len(heading_words & line_words)
if match_count > best_match_score:
best_match_score = match_count
matched_header_line = line_text.strip()
if matched_header_line:
print(f"✅ Best match: '{matched_header_line}' with score {best_match_score}")
else:
print("❌ No suitable match found.")
return
else:
matched_header_line = target_header # Exact match
# matched_header_line = target_header
matched_header_font_size = most_common_font_size
collecting = False
collected_lines = []
page_highlights = {}
current_bbox = {}
last_y1s = {}
mainHeader=''
subHeader=''
matched_header_line_norm = normalize_text(matched_header_line)
color = color_manager.get_next_color()
for page_num in range(len(doc)):
if page_num in toc_pages:
continue
page = doc.load_page(page_num)
page_height = page.rect.height
blocks = page.get_text("dict")["blocks"]
for block in blocks:
lines = block.get("lines", [])
i = 0
while i < len(lines):
spans = lines[i].get("spans", [])
if not spans:
i += 1
continue
y0 = spans[0]["bbox"][1]
y1 = spans[0]["bbox"][3]
if y0 < top_margin or y1 > (page_height - bottom_margin):
i += 1
continue
# print(line_text)
line_text = get_spaced_text_from_spans(spans).lower()
line_text_norm = normalize_text(line_text)
if i + 1 < len(lines):
next_spans = lines[i + 1].get("spans", [])
next_line_text = get_spaced_text_from_spans(next_spans).lower()
combined_line = (line_text + " " + next_line_text).strip()
combined_line_norm = normalize_text(combined_line)
else:
combined_line = line_text
combined_line_norm = line_text_norm
# if not done and not collecting:
if not done and not collecting:
for span in spans:
if len(normalize_text(span['text'])) > 1:
if is_header(span, most_common_font_size, most_common_color, most_common_font):
for header in headers:
header_text, header_size, header_page, header_y = header # 4 elements now!
# Check if combined_line_norm is inside header text
if combined_line_norm in header_text:
# Also check that the Y position is close (for example, within 5 pixels)
# if abs(span['bbox'][1] - header_y) < 1:
print('comb:,',combined_line_norm)
if header_size == mainHeaderFontSize:
mainHeader=find_full_line_in_toc(doc, toc_pages, combined_line_norm)
print('main:', mainHeader)
elif header_size == subHeaderFontSize:
subHeader = combined_line_norm
print('sub:', subHeader)
# Start collecting if we find the target header
if matched_header_line_norm in combined_line_norm and not collecting:
if any(is_header(span, most_common_font_size, most_common_color, most_common_font) for span in spans):
collecting = True
header_font_sizes = [span["size"] for span in spans if is_header(span, most_common_font_size, most_common_color, most_common_font)]
if header_font_sizes:
matched_header_font_size = max(header_font_sizes)
print(f"📥 Start collecting after header: {combined_line} (Font size: {matched_header_font_size})")
pageNumberFound = page_num +1
# Collect the header line text and bbox too!
collected_lines.append(line_text)
valid_spans = [span for span in spans if span.get("bbox")]
if valid_spans:
x0s = [span["bbox"][0] for span in valid_spans]
x1s = [span["bbox"][2] for span in valid_spans]
y0s = [span["bbox"][1] for span in valid_spans]
y1s = [span["bbox"][3] for span in valid_spans]
left = int(x0s[0])
top = int(y0s[0])
print(left,type(left),top,type(top))
header_bbox = [min(x0s), min(y0s), max(x1s), max(y1s)]
if page_num in current_bbox:
cb = current_bbox[page_num]
current_bbox[page_num] = [
min(cb[0], header_bbox[0]),
min(cb[1], header_bbox[1]),
max(cb[2], header_bbox[2]),
max(cb[3], header_bbox[3])
]
else:
current_bbox[page_num] = header_bbox
last_y1s[page_num] = header_bbox[3]
i += 2
continue
if collecting:
norm_line = normalize_text(line_text)
norm_combined = normalize_text(combined_line)
# 🧠 Skip URL-like lines from being considered headers
if re.match(r'https?://\S+|www\.\S+', norm_line):
line_is_header = False
else:
line_is_header = any(is_header(span, most_common_font_size, most_common_color, most_common_font) for span in spans)
if line_is_header:
header_font_size = max(span["size"] for span in spans)
is_probably_real_header = (
header_font_size >= matched_header_font_size and
is_header(spans[0], most_common_font_size, most_common_color, most_common_font) and
len(line_text.strip()) > 2
)
if (norm_line != matched_header_line_norm and
norm_combined != matched_header_line_norm and
is_probably_real_header):
print(f"🛑 Stop at header with same or larger font: '{line_text}' ({header_font_size}{matched_header_font_size})")
collecting = False
done=True
result_text = (matched_header_line + "\n" + "\n".join(collected_lines)).strip().lower()
print("\n📄 Final collected section (early return):\n" , mainHeader,subHeader)
print(result_text)
for page_num, bbox in current_bbox.items():
# update y1 to stop exactly at last_y1
bbox[3] = last_y1s.get(page_num, bbox[3])
page_highlights[page_num] = bbox
highlight_boxes(doc, page_highlights,color)
zoom = 200
zoom_str = f"{zoom},{left},{top}"
print('zoooom',zoom_str)
params = {
'pdfLink': pdf_path, # Your PDF link
'keyword': heading_to_search, # Your keyword (could be a string or list)
}
# URL encode each parameter
encoded_params = {key: urllib.parse.quote(value, safe='') for key, value in params.items()}
# Construct the final encoded link
encoded_link = '&'.join([f"{key}={value}" for key, value in encoded_params.items()])
# Correctly construct the final URL with page and zoom
zoom_str = f"{zoom},{left},{top}"
final_url = f"{baselink}{encoded_link}#page={str(pageNumberFound)}&zoom={zoom_str}"
print(final_url)
# Get current date and time
now = datetime.now()
# Format the output
formatted_time = now.strftime("%d/%m/%Y %I:%M:%S %p")
if mainHeader:
data_entry = {
"NBSLink": final_url,
"Subject": 'Markup (initial)',
"Page": str(pageNumberFound),
"Author": "ADR",
"Creation Date": formatted_time,
"Layer": "Initial",
"Code": heading_to_search,
"head above 1": mainHeader,
"head above 2": subHeader
}
data_list_JSON.append(data_entry)
# Convert list to JSON
print('heree')
# json_output = json.dumps(data_list_JSON, indent=4)
# return result_text
collected_lines.append(line_text)
valid_spans = [span for span in spans if span.get("bbox")]
if valid_spans:
x0s = [span["bbox"][0] for span in valid_spans]
x1s = [span["bbox"][2] for span in valid_spans]
y0s = [span["bbox"][1] for span in valid_spans]
y1s = [span["bbox"][3] for span in valid_spans]
line_bbox = [min(x0s), min(y0s), max(x1s), max(y1s)]
if page_num in current_bbox:
cb = current_bbox[page_num]
current_bbox[page_num] = [
min(cb[0], line_bbox[0]),
min(cb[1], line_bbox[1]),
max(cb[2], line_bbox[2]),
max(cb[3], line_bbox[3])
]
else:
current_bbox[page_num] = line_bbox
last_y1s[page_num] = line_bbox[3]
i += 1
# doc.save("highlighted_output.pdf", garbage=4, deflate=True)
result_text = (matched_header_line + "\n" + "\n".join(collected_lines)).strip().lower()
print("\n📄 Final collected section:\n")
pdf_bytes = BytesIO()
doc.save(pdf_bytes)
print('aa')
print('JSONN',data_list_JSON)
return pdf_bytes.getvalue(), doc , df, data_list_JSON