neuralworm's picture
better json output
8902f5e
raw
history blame
25.6 kB
#TODO: Quran results have numbers
import logging
logger = logging.getLogger(__name__)
logging.basicConfig(level=logging.INFO)
import gradio as gr
import torah
import bible
import quran
import hindu
import tripitaka
from utils import number_to_ordinal_word, custom_normalize, date_to_words, translate_date_to_words
from gematria import calculate_gematria, strip_diacritics
import pandas as pd
from deep_translator import GoogleTranslator
from gradio_calendar import Calendar
from datetime import datetime, timedelta
import math
import json
import re
import sqlite3
from collections import defaultdict
from typing import List, Tuple
import rich
from fuzzywuzzy import fuzz
import calendar
import translation_utils
import hashlib
translation_utils.create_translation_table()
# Create a translator instance *once* globally
translator = GoogleTranslator(source='auto', target='auto')
LANGUAGES_SUPPORTED = translator.get_supported_languages(as_dict=True) # Corrected dictionary name
LANGUAGE_CODE_MAP = LANGUAGES_SUPPORTED # Use deep_translator's mapping directly
# --- Constants ---
DATABASE_FILE = 'gematria.db'
MAX_PHRASE_LENGTH_LIMIT = 20
ELS_CACHE_DB = "els_cache.db"
DATABASE_TIMEOUT = 60
# --- Database Initialization ---
def initialize_database():
global conn
conn = sqlite3.connect(DATABASE_FILE)
cursor = conn.cursor()
cursor.execute('''
CREATE TABLE IF NOT EXISTS results (
gematria_sum INTEGER,
words TEXT,
translation TEXT,
book TEXT,
chapter INTEGER,
verse INTEGER,
phrase_length INTEGER,
word_position TEXT,
PRIMARY KEY (gematria_sum, words, book, chapter, verse, word_position)
)
''')
cursor.execute('''
CREATE INDEX IF NOT EXISTS idx_results_gematria
ON results (gematria_sum)
''')
cursor.execute('''
CREATE TABLE IF NOT EXISTS processed_books (
book TEXT PRIMARY KEY,
max_phrase_length INTEGER
)
''')
conn.commit()
# --- Initialize Database ---
initialize_database()
# --- ELS Cache Functions ---
def create_els_cache_table():
with sqlite3.connect(ELS_CACHE_DB) as conn:
conn.execute('''
CREATE TABLE IF NOT EXISTS els_cache (
query_hash TEXT PRIMARY KEY,
results TEXT
)
''')
def get_query_hash(func, *args, **kwargs):
key = (func.__name__, args, tuple(sorted(kwargs.items())))
return hashlib.sha256(json.dumps(key).encode()).hexdigest()
def cached_process_json_files(func, *args, **kwargs):
query_hash = get_query_hash(func, *args, **kwargs)
try:
with sqlite3.connect(ELS_CACHE_DB, timeout=DATABASE_TIMEOUT) as conn:
cursor = conn.cursor()
cursor.execute("SELECT results FROM els_cache WHERE query_hash = ?", (query_hash,))
result = cursor.fetchone()
if result:
logger.info(f"Cache hit for query: {query_hash}")
return json.loads(result[0])
except sqlite3.Error as e:
logger.error(f"Database error checking cache: {e}")
logger.info(f"Cache miss for query: {query_hash}")
results = func(*args, **kwargs)
try:
with sqlite3.connect(ELS_CACHE_DB, timeout=DATABASE_TIMEOUT) as conn:
cursor = conn.cursor()
cursor.execute("INSERT INTO els_cache (query_hash, results) VALUES (?, ?)", (query_hash, json.dumps(results)))
conn.commit()
except sqlite3.Error as e:
logger.error(f"Database error caching results: {e}")
return results
# --- Helper Functions (from Network app.py) ---
def flatten_text(text: List) -> str:
if isinstance(text, list):
return " ".join(flatten_text(item) if isinstance(item, list) else item for item in text)
return text
def search_gematria_in_db(gematria_sum: int, max_words: int) -> List[Tuple[str, str, int, int, int, str]]:
global conn
with sqlite3.connect(DATABASE_FILE) as conn:
cursor = conn.cursor()
cursor.execute('''
SELECT words, book, chapter, verse, phrase_length, word_position
FROM results
WHERE gematria_sum = ? AND phrase_length <= ?
''', (gematria_sum, max_words))
results = cursor.fetchall()
return results
def get_most_frequent_phrase(results):
phrase_counts = defaultdict(int)
for words, book, chapter, verse, phrase_length, word_position in results:
phrase_counts[words] += 1
most_frequent_phrase = max(phrase_counts, key=phrase_counts.get) if phrase_counts else None # Handle empty results
return most_frequent_phrase
# --- Functions from BOS app.py ---
def create_language_dropdown(label, default_value='English', show_label=True): # Default value must be in LANGUAGE_CODE_MAP
return gr.Dropdown(
choices=list(LANGUAGE_CODE_MAP.keys()), # Correct choices
label=label,
value=default_value,
show_label=show_label
)
def calculate_gematria_sum(text, date_words):
if text or date_words:
combined_input = f"{text} {date_words}"
logger.info(f"searching for input: {combined_input}")
numbers = re.findall(r'\d+', combined_input)
text_without_numbers = re.sub(r'\d+', '', combined_input)
number_sum = sum(int(number) for number in numbers)
text_gematria = calculate_gematria(strip_diacritics(text_without_numbers))
total_sum = text_gematria + number_sum
return total_sum
else:
return None
def perform_els_search(step, rounds_combination, tlang, strip_spaces, strip_in_braces, strip_diacritics_chk, include_torah, include_bible, include_quran, include_hindu, include_tripitaka):
if step == 0 or rounds_combination == "0,0":
return None
results = {}
length = 0
selected_language_long = tlang # From the Gradio dropdown (long form)
tlang = LANGUAGES_SUPPORTED.get(selected_language_long) #Get the short code.
if tlang is None: # Handle unsupported languages
tlang = "en"
logger.warning(f"Unsupported language selected: {selected_language_long}. Defaulting to English (en).")
if include_torah:
logger.debug(f"Arguments for Torah: {(1, 39, step, rounds_combination, length, tlang, strip_spaces, strip_in_braces, strip_diacritics_chk)}")
results["Torah"] = cached_process_json_files(torah.process_json_files, 1, 39, step, rounds_combination, length, tlang, strip_spaces, strip_in_braces, strip_diacritics_chk)
else:
results["Torah"] = []
if include_bible:
results["Bible"] = cached_process_json_files(bible.process_json_files, 40, 66, step, rounds_combination, length, tlang, strip_spaces, strip_in_braces, strip_diacritics_chk)
else:
results["Bible"] = []
if include_quran:
results["Quran"] = cached_process_json_files(quran.process_json_files, 1, 114, step, rounds_combination, length, tlang, strip_spaces, strip_in_braces, strip_diacritics_chk)
else:
results["Quran"] = []
if include_hindu:
results["Rig Veda"] = cached_process_json_files(hindu.process_json_files, 1, 10, step, rounds_combination, length, tlang, False, strip_in_braces, strip_diacritics_chk)
else:
results["Rig Veda"] = []
if include_tripitaka:
results["Tripitaka"] = cached_process_json_files(tripitaka.process_json_files, 1, 52, step, rounds_combination, length, tlang, strip_spaces, strip_in_braces, strip_diacritics_chk)
else:
results["Tripitaka"] = []
return results
def add_24h_projection(results_dict): #Now takes a dictionary of results
for book_name, results in results_dict.items(): # Iterate per book
num_results = len(results)
if num_results > 0:
time_interval = timedelta(minutes=24 * 60 / num_results)
current_time = datetime.min.time()
for i in range(num_results):
next_time = (datetime.combine(datetime.min, current_time) + time_interval).time()
time_range_str = f"{current_time.strftime('%H:%M')}-{next_time.strftime('%H:%M')}"
results[i]['24h Projection'] = time_range_str
current_time = next_time
return results_dict
def add_monthly_projection(results_dict, selected_date):
if selected_date is None:
return results_dict # Return if no date is selected
for book_name, results in results_dict.items(): # Iterate per book
num_results = len(results)
if num_results > 0:
days_in_month = calendar.monthrange(selected_date.year, selected_date.month)[1]
total_seconds = (days_in_month - 1) * 24 * 3600
seconds_interval = total_seconds / num_results
start_datetime = datetime(selected_date.year, selected_date.month, 1)
current_datetime = start_datetime
for i in range(num_results):
next_datetime = current_datetime + timedelta(seconds=seconds_interval)
current_date = current_datetime.date() # Moved assignment inside loop
next_date = next_datetime.date()
date_range_str = f"{current_date.strftime('%h %d')} - {next_date.strftime('%h %d')}"
results[i]['Monthly Projection'] = date_range_str
current_datetime = next_datetime # Add this
current_date = next_datetime.date() # Add this too
return results_dict
def add_yearly_projection(results_dict, selected_date): #Correct name, handle dictionary input
if selected_date is None:
return results_dict # Return if no date is selected
for book_name, results in results_dict.items(): # Iterate per book
num_results = len(results)
if num_results > 0:
days_in_year = 366 if calendar.isleap(selected_date.year) else 365
total_seconds = (days_in_year - 1) * 24 * 3600
seconds_interval = total_seconds / num_results
start_datetime = datetime(selected_date.year, 1, 1)
current_datetime = start_datetime
for i in range(num_results):
next_datetime = current_datetime + timedelta(seconds=seconds_interval)
current_date = current_datetime.date() # Move assignment inside loop
next_date = next_datetime.date()
date_range_str = f"{current_date.strftime('%b %d')} - {next_date.strftime('%b %d')}"
results[i]['Yearly Projection'] = date_range_str
current_datetime = next_datetime # Update current datetime for next iteration
return results_dict
def sort_results(results):
def parse_time(time_str):
try:
hours, minutes = map(int, time_str.split(':'))
return hours * 60 + minutes # Convert to total minutes
except ValueError:
return 24 * 60 # Sort invalid times to the end
return sorted(results, key=lambda x: (
parse_time(x.get('24h Projection', '23:59').split('-')[0]), # Sort by start time first
parse_time(x.get('24h Projection', '23:59').split('-')[1]) # Then by end time
))
# --- Main Gradio App ---
with gr.Blocks() as app:
with gr.Column():
with gr.Row():
tlang = create_language_dropdown("Target Language for Result Translation", default_value='english')
selected_date = Calendar(type="datetime", label="Date to investigate (optional)", info="Pick a date from the calendar")
use_day = gr.Checkbox(label="Use Day", info="Check to include day in search", value=True)
use_month = gr.Checkbox(label="Use Month", info="Check to include month in search", value=True)
use_year = gr.Checkbox(label="Use Year", info="Check to include year in search", value=True)
date_language_input = create_language_dropdown("Language of the person/topic (optional) (Date Word Language)", default_value='english')
with gr.Row():
gematria_text = gr.Textbox(label="Name and/or Topic (required)", value="Hans Albert Einstein Mileva Marity-Einstein")
date_words_output = gr.Textbox(label="Date in Words Translated (optional)")
gematria_result = gr.Number(label="Journal Sum")
#with gr.Row():
with gr.Row():
step = gr.Number(label="Jump Width (Steps) for ELS")
float_step = gr.Number(visible=False, value=1)
half_step_btn = gr.Button("Steps / 2")
double_step_btn = gr.Button("Steps * 2")
with gr.Column():
round_x = gr.Number(label="Round (1)", value=1)
round_y = gr.Number(label="Round (2)", value=-1)
rounds_combination = gr.Textbox(label="Combined Rounds", value="1,-1")
with gr.Row():
include_torah_chk = gr.Checkbox(label="Include Torah", value=True)
include_bible_chk = gr.Checkbox(label="Include Bible", value=True)
include_quran_chk = gr.Checkbox(label="Include Quran", value=True)
include_hindu_chk = gr.Checkbox(label="Include Rigveda", value=True)
include_tripitaka_chk = gr.Checkbox(label="Include Tripitaka", value=True)
merge_results_chk = gr.Checkbox(label="Merge Results (Torah-Bible-Quran)", value=True)
strip_spaces = gr.Checkbox(label="Strip Spaces from Books", value=True)
strip_in_braces = gr.Checkbox(label="Strip Text in Braces from Books", value=True)
strip_diacritics_chk = gr.Checkbox(label="Strip Diacritics from Books", value=True)
translate_btn = gr.Button("Search with ELS")
# --- Output Components ---
markdown_output = gr.Dataframe(label="ELS Results")
most_frequent_phrase_output = gr.Textbox(label="Most Frequent Phrase in Network Search")
json_output = gr.JSON(label="JSON Output")
# --- Event Handlers ---
def update_date_words(selected_date, date_language_input, use_day, use_month, use_year):
if selected_date is None:
return ""
if not use_year and not use_month and not use_day:
return translate_date_to_words(selected_date, date_language_input)
year = selected_date.year if use_year else None
month = selected_date.month if use_month else None
day = selected_date.day if use_day else None
if year is not None and month is not None and day is not None:
date_obj = selected_date
elif year is not None and month is not None:
date_obj = str(f"{year}-{month}")
elif year is not None:
date_obj = str(f"{year}")
else: # Return empty string if no date components are selected
return ""
date_in_words = date_to_words(date_obj)
translator = GoogleTranslator(source='auto', target=date_language_input)
translated_date_words = translator.translate(date_in_words)
return custom_normalize(translated_date_words)
def update_journal_sum(gematria_text, date_words_output):
sum_value = calculate_gematria_sum(gematria_text, date_words_output)
return sum_value, sum_value, sum_value
def update_rounds_combination(round_x, round_y):
return f"{int(round_x)},{int(round_y)}"
def update_step_half(float_step):
new_step = math.ceil(float_step / 2)
return new_step, float_step / 2
def update_step_double(float_step):
new_step = math.ceil(float_step * 2)
return new_step, float_step * 2
def find_closest_phrase(target_phrase, phrases):
best_match = None
best_score = 0
logging.debug(f"Target phrase for similarity search: {target_phrase}") # Log target phrase
for phrase, _, _, _, _, _ in phrases:
word_length_diff = abs(len(target_phrase.split()) - len(phrase.split()))
similarity_score = fuzz.ratio(target_phrase, phrase)
combined_score = similarity_score - word_length_diff
logging.debug(f"Comparing with phrase: {phrase}") # Log each phrase being compared
logging.debug(
f"Word Length Difference: {word_length_diff}, Similarity Score: {similarity_score}, Combined Score: {combined_score}") # Log scores
if combined_score > best_score:
best_score = combined_score
best_match = phrase
logging.debug(f"Closest phrase found: {best_match} with score: {best_score}") # Log the best match
return best_match
def perform_search(step, rounds_combination, tlang, strip_spaces, strip_in_braces, strip_diacritics_chk, include_torah, include_bible, include_quran, include_hindu, include_tripitaka, gematria_text, date_words_output, selected_date):
# Inside perform_search
els_results = perform_els_search(step, rounds_combination, tlang, strip_spaces, strip_in_braces,
strip_diacritics_chk, include_torah, include_bible, include_quran,
include_hindu,
include_tripitaka)
# --- Network Search Integration ---
most_frequent_phrases = {}
combined_and_sorted_results = [] # Combined list to hold all results
for book_name, book_results in els_results.items():
if book_results: # Add this check to ensure book_results is not empty
most_frequent_phrases[book_name] = "" # Default value
for result in book_results:
try:
gematria_sum = calculate_gematria(result['result_text']) # Calculate gematria
max_words = len(result['result_text'].split())
matching_phrases = search_gematria_in_db(gematria_sum, max_words)
max_words_limit = 20
while not matching_phrases and max_words < max_words_limit: # Increase max_words for more results
max_words += 1
matching_phrases = search_gematria_in_db(gematria_sum, max_words)
if matching_phrases:
most_frequent_phrase = get_most_frequent_phrase(matching_phrases)
most_frequent_phrases[book_name] = most_frequent_phrase
else:
closest_phrase = find_closest_phrase(result['result_text'],
search_gematria_in_db(gematria_sum, max_words_limit))
most_frequent_phrases[
book_name] = closest_phrase or "" # Update most frequent phrases even if no phrase found
result['Most Frequent Phrase'] = most_frequent_phrases[book_name]
if 'book' in result:
if isinstance(result['book'], int): # Torah, Bible, Quran case
result['book'] = f"{book_name} {result['book']}."
combined_and_sorted_results.append(result)
except KeyError as e:
print(f"DEBUG: KeyError - Key '{e.args[0]}' not found in result. Skipping this result.")
continue
# --- Batch Translation ---
selected_language_long = tlang
tlang_short = LANGUAGES_SUPPORTED.get(selected_language_long)
if tlang_short is None:
tlang_short = "en"
logger.warning(f"Unsupported language selected: {selected_language_long}. Defaulting to English (en).")
# Prepare lists for batch translation, including source language
phrases_to_translate = []
phrases_source_langs = [] # Source languages for phrases
results_to_translate = []
results_source_langs = [] # Source languages for results
for result in combined_and_sorted_results:
phrases_to_translate.append(result.get('Most Frequent Phrase', ''))
phrases_source_langs.append(result.get("source_language", "auto"))
results_to_translate.append(result.get('result_text', ''))
results_source_langs.append(result.get("source_language", "auto"))
translated_phrases = translation_utils.batch_translate(phrases_to_translate, tlang_short, phrases_source_langs)
translated_result_texts = translation_utils.batch_translate(results_to_translate, tlang_short, results_source_langs)
for i, result in enumerate(combined_and_sorted_results):
result['translated_text'] = translated_result_texts.get(results_to_translate[i], None)
result['Translated Most Frequent Phrase'] = translated_phrases.get(phrases_to_translate[i], None)
# Time Projections (using els_results dictionary)
updated_els_results = add_24h_projection(els_results) # Use original els_results dictionary
updated_els_results = add_monthly_projection(updated_els_results, selected_date) # Call correct functions with correct params
updated_els_results = add_yearly_projection(updated_els_results, selected_date)
combined_and_sorted_results = []
for book_results in updated_els_results.values(): # Combine results for dataframe and json
combined_and_sorted_results.extend(book_results)
combined_and_sorted_results = sort_results(combined_and_sorted_results) # sort combined results
df = pd.DataFrame(combined_and_sorted_results)
df.index = range(1, len(df) + 1)
df.reset_index(inplace=True)
df.rename(columns={'index': 'Result Number'}, inplace=True)
for i, result in enumerate(combined_and_sorted_results): # Iterate through the combined list
result['Result Number'] = i + 1
search_config = {
"step": step,
"rounds_combination": rounds_combination,
"target_language": tlang,
"strip_spaces": strip_spaces,
"strip_in_braces": strip_in_braces,
"strip_diacritics": strip_diacritics_chk,
"include_torah": include_torah,
"include_bible": include_bible,
"include_quran": include_quran,
"include_hindu": include_hindu,
"include_tripitaka": include_tripitaka,
"gematria_text": gematria_text,
"date_words": date_words_output
}
output_data = {
"search_configuration": search_config,
"results": combined_and_sorted_results # Use the combined list here
}
json_data = output_data
# --- Return results ---
combined_most_frequent = "\n".join(
f"{book}: {phrase}" for book, phrase in most_frequent_phrases.items()) # Combine phrases
return df, combined_most_frequent, json_data
# --- Event Triggers ---
round_x.change(update_rounds_combination, inputs=[round_x, round_y], outputs=rounds_combination)
round_y.change(update_rounds_combination, inputs=[round_x, round_y], outputs=rounds_combination)
selected_date.change(update_date_words, inputs=[selected_date, date_language_input, use_day, use_month, use_year], outputs=[date_words_output])
date_language_input.change(update_date_words, inputs=[selected_date, date_language_input, use_day, use_month, use_year], outputs=[date_words_output])
gematria_text.change(update_journal_sum, inputs=[gematria_text, date_words_output], outputs=[gematria_result, step, float_step])
date_words_output.change(update_journal_sum, inputs=[gematria_text, date_words_output], outputs=[gematria_result, step, float_step])
half_step_btn.click(update_step_half, inputs=[float_step], outputs=[step, float_step])
double_step_btn.click(update_step_double, inputs=[float_step], outputs=[step, float_step])
translate_btn.click(
perform_search,
inputs=[step, rounds_combination, tlang, strip_spaces, strip_in_braces, strip_diacritics_chk, include_torah_chk, include_bible_chk, include_quran_chk, include_hindu_chk, include_tripitaka_chk, gematria_text, date_words_output, selected_date],
outputs=[markdown_output, most_frequent_phrase_output, json_output]
)
app.load(
update_date_words,
inputs=[selected_date, date_language_input, use_day, use_month, use_year], # Include all 5 inputs
outputs=[date_words_output]
)
use_day.change(
update_date_words,
inputs=[selected_date, date_language_input, use_day, use_month, use_year],
outputs=[date_words_output]
)
use_month.change(
update_date_words,
inputs=[selected_date, date_language_input, use_day, use_month, use_year],
outputs=[date_words_output]
)
use_year.change(
update_date_words,
inputs=[selected_date, date_language_input, use_day, use_month, use_year],
outputs=[date_words_output]
)
def checkbox_behavior(use_day_value, use_month_value):
if use_day_value: # Tick month and year automatically when day is ticked.
return True, True
return use_month_value, True # return month value unchanged and automatically tick year if month is checked
use_day.change(checkbox_behavior, inputs=[use_day, use_month], outputs=[use_month, use_year])
use_month.change(checkbox_behavior, inputs=[use_day, use_month], outputs=[use_month, use_year]) #No need for use_day here, day won't be changed by month
if __name__ == "__main__":
app.launch(share=False)