context_analyzer.py

import re

# Basic regex patterns - these can be significantly improved for accuracy and internationalization.

# Regex for emails: Handles most common email formats.
# Allows for longer TLDs and stricter local part.
EMAIL_REGEX = r"[a-zA-Z0-9!#$%&'*+\-/=?^_`{|}~.;]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,20}"

# Regex for URLs: Improved to better capture various URL forms, including those without http/s but with www,
# and basic path/query/fragment. Still not exhaustive for all edge cases like IP addresses or ports.
URL_REGEX = r"(?:(?:https?|ftp):\/\/|www\.)[-a-zA-Z0-9@:%._\+~#=]{1,256}\.[a-zA-Z0-9()]{1,6}\b(?:[-a-zA-Z0-9()@:%_\+.~#?&//=]*)"

# Regex for phone numbers: Expanded to include more variations, still primarily focused on common structures.
# This is notoriously difficult. For truly robust parsing, a dedicated library is better.
# This pattern attempts to match:
# - Optional international prefix like +1, +49, 0049
# - Optional area codes in parentheses
# - Various separators: space, dash, dot
# - Common lengths for local numbers (e.g., 7 to 10 digits after area code)
PHONE_REGEX = r"(?:\+?\d{1,4}[-.\s]?)?(?:\(?\d{1,5}\)?[-.\s]?)?\d{3,5}[-.\s]?\d{3,5}(?:[-.\s]?\d{1,5})?"
# Add a more specific German phone number pattern to catch those first.
# Example: 0171 1234567 or +49 30 1234567 or (030) 123456-78
GERMAN_PHONE_REGEX = r"(?:(?:\(?(?:0049|0)\)?\s?\d{1,5})|(?:\+49\s?\d{1,5}))(?:(?:[\s.-]?\d{2,}){2,}|[\s.-]?\d{4,12})"


# Regex for dates: Expanded for more common numeric formats and separators.
# Still does not validate day/month ranges.
# Catches YYYY-MM-DD, YYYY/MM/DD, YYYY.MM.DD
# Catches DD.MM.YYYY, DD/MM/YYYY, DD-MM-YYYY
# Catches MM/DD/YYYY (common US)
DATE_REGEX = r"""
    \b(
        \d{4}[-/.]\d{1,2}[-/.]\d{1,2} | # YYYY-MM-DD and variations
        \d{1,2}[-/.]\d{1,2}[-/.]\d{4}   # DD.MM.YYYY and MM/DD/YYYY and variations
    )\b
"""


def find_emails(text: str) -> list[str]:
    """Finds all email addresses in the given text."""
    if not text: return []
    return list(set(re.findall(EMAIL_REGEX, text, re.IGNORECASE))) # Use set to get unique emails

def find_urls(text: str) -> list[str]:
    """Finds all URLs in the given text."""
    if not text: return []
    # A common issue is OCR splitting URLs with spaces if they wrap lines.
    # This regex won't handle that; pre-processing of OCR text might be needed.
    return list(set(re.findall(URL_REGEX, text, re.IGNORECASE)))

def find_phone_numbers(text: str) -> list[str]:
    """Finds all phone numbers (based on improved patterns) in the given text."""
    if not text: return []
    # Try German specific first, then the more general one.
    # Use a set to avoid duplicates if patterns overlap.
    found_numbers = set()
    # German numbers might have spaces that the general regex splits, so clean them.
    # For German regex, remove spaces within potential numbers before matching.
    # This is a heuristic.
    text_no_space_for_german_phones = re.sub(r"(\d)\s+(\d)", r"\1\2", text)

    german_matches = re.findall(GERMAN_PHONE_REGEX, text_no_space_for_german_phones)
    for num in german_matches:
        # Basic cleanup: remove common separators to somewhat normalize
        cleaned_num = re.sub(r"[\s().\-/]", "", num)
        if len(cleaned_num) >= 7: # Arbitrary minimum length for a "valid" cleaned number
             found_numbers.add(num.strip()) # Add original formatting

    general_matches = re.findall(PHONE_REGEX, text)
    for num in general_matches:
        cleaned_num = re.sub(r"[\s().\-/]", "", num)
        if len(cleaned_num) >= 7:
            found_numbers.add(num.strip())

    return sorted(list(found_numbers), key=len, reverse=True) # Prefer longer matches

def find_dates(text: str) -> list[str]:
    """Finds all dates (based on improved patterns) in the given text."""
    if not text: return []
    # The re.VERBOSE flag allows for comments and cleaner layout in the DATE_REGEX.
    return list(set(re.findall(DATE_REGEX, text, re.VERBOSE)))

def analyze_text_for_contextual_items(text: str) -> dict:
    """
    Analyzes text and returns a dictionary of found items.
    Keys are item types (e.g., 'emails', 'urls'), values are lists of found strings.
    """
    if not text or not text.strip():
        return {}

    found_items = {
        "emails": find_emails(text),
        "urls": find_urls(text),
        "phones": find_phone_numbers(text),
        "dates": find_dates(text),
    }
    # Filter out empty lists
    return {k: v for k, v in found_items.items() if v}


if __name__ == '__main__':
    sample_text_1 = """
    Hello, please contact me at test.user@example.com or my_other.email@sub.example.co.uk.
    Visit our website http://www.example.com or https://another-example.com/path?query=true.
    Call me on (123) 456-7890 or +49 123 45678. My old number was 555-123-4567.
    Important dates: 2023-12-25, 01/15/2024, and 31.07.2023. Not a date: 1234.
    This is just a regular number 12345 and a price $19.99.
    """

    sample_text_2 = "No special items here. Just plain text."

    sample_text_3 = "My website is www.test-site.com."

    print("--- Analyzing Sample Text 1 ---")
    items1 = analyze_text_for_contextual_items(sample_text_1)
    for item_type, values in items1.items():
        print(f"{item_type.capitalize()}: {values}")

    print("\n--- Analyzing Sample Text 2 ---")
    items2 = analyze_text_for_contextual_items(sample_text_2)
    if not items2:
        print("No contextual items found.")
    else:
        for item_type, values in items2.items():
            print(f"{item_type.capitalize()}: {values}")

    print("\n--- Analyzing Sample Text 3 (www only) ---")
    items3 = analyze_text_for_contextual_items(sample_text_3)
    for item_type, values in items3.items():
        print(f"{item_type.capitalize()}: {values}")

    print("\n--- Testing individual finders ---")
    print(f"Emails in 'contact support@example.org': {find_emails('contact support@example.org')}")
    print(f"URLs in 'go to example.com': {find_urls('go to example.com')}") # Will not find without http/www by current regex
    print(f"URLs in 'go to www.example.com': {find_urls('go to www.example.com')}")
    print(f"Phones in 'Tel: 1234567890': {find_phone_numbers('Tel: 1234567890')}")
    print(f"Dates in 'Meeting on 10/20/2025': {find_dates('Meeting on 10/20/2025')}")