more improvements and bugfixing

This commit is contained in:
2025-10-25 23:18:52 -07:00
parent 64939ac1b5
commit 0106f0b9ac
6 changed files with 442 additions and 23 deletions
+172 -17
View File
@@ -113,6 +113,10 @@ for group in theological_synonym_groups:
# Map each word to other words in its group
theological_synonyms_map[word] = [w for w in group if w != word]
# Load concept mappings for common theological terms not in KJV
concept_mappings = json.load(open("data/concept_mappings.json", "r"))
concept_mappings_dict = {mapping["query"].lower(): mapping for mapping in concept_mappings}
cached: dict = {
"verse_ids": [i[0] for i in kjv_cur.execute("SELECT id FROM fts_kjv;").fetchall()],
"book_names": [
@@ -123,6 +127,7 @@ cached: dict = {
"json_dictionary": json.load(open("data/1828_Webster_KJV.json", "r")),
"grammatical_variants": grammatical_variants_map,
"theological_synonyms": theological_synonyms_map,
"concept_mappings": concept_mappings_dict,
"html_response": open("html/response.html", "r").read(),
"html_navbar": open("html/navbar.html", "r").read(),
"css_navbar": open("styles/header.css", "r").read(),
@@ -470,34 +475,55 @@ def parse_query(query_override: str = None):
return send_file("html/find.html")
if query.endswith(","):
query = query[:-1]
result = show(kw=query, human_readable=human_readable)
if not hasattr(result, "__len__"):
logging.debug(f"parse_query.show = '{result}'")
return result
if len(result) > 17:
return result
result = search()
# Check if query contains both letters and digits (required for reference parsing)
# If only numbers or only letters, skip reference parsing and go to keyword search
has_letters = any(char.isalpha() for char in query)
has_digits = any(char.isdigit() for char in query)
if not (has_letters and has_digits):
result = search()
else:
result = show(kw=query, human_readable=human_readable)
if not hasattr(result, "__len__"):
logging.debug(f"parse_query.show = '{result}'")
return result
if len(result) > 17:
return result
# Reference parsing failed, fall back to keyword search
result = search()
# Remove the closing </div> from search() so we can add dictionary content inside .responses
if result.endswith("</div>"):
result = result[:-6] # Remove last </div>
result += "<style>" + cached["css_collapsible"] + "</style>"
result += "<div class='dictionary-section'><h3>1828 Webster Definitions</h3><p>Click on the word to get the definition.</p></div>"
# Check if we have any dictionary definitions before showing the section
split_query = query.replace(",", " ").split(" ")
if len(split_query) > 20:
return "Your Query is too long. 20 Words or less."
definitions_html = ""
has_definitions = False
for word in split_query:
definition = define(word)
if not definition:
continue
result += (
has_definitions = True
definitions_html += (
"<button type='button' class='collapsible'>"
+ word.capitalize()
+ "</button>"
)
result += "<div class='definition'>" + definition + "</div>"
definitions_html += "<div class='definition'>" + definition + "</div>"
logging.debug(f"parse_query.define = '{word}'")
result += "<script src='/js/collapsible.js'></script>"
# Only add the dictionary section if we found at least one definition
if has_definitions:
result += "<style>" + cached["css_collapsible"] + "</style>"
result += "<div class='dictionary-section'><h3>1828 Webster Definitions</h3><p>Click on the word to get the definition.</p></div>"
result += definitions_html
result += "<script src='/js/collapsible.js'></script>"
result += "</div>"
return result
@@ -663,7 +689,7 @@ def response_rich(scripture):
def scripture_response(scriptures, human_readable=False):
response = []
arg_view = request.args.get("view", "json")
arg_view = request.args.get("view", "rich")
logging.debug(f"search ->requested_format = {arg_view}")
if arg_view == "plain":
for scripture in scriptures:
@@ -789,7 +815,88 @@ def search():
return send_file("html/search.html")
else:
return send_file("html/search.html")
# Check if this is a common theological concept not in KJV
concept_mapping = cached["concept_mappings"].get(cleaned_keywords.lower(), None)
concept_notice = None
if concept_mapping:
# Build hyperlinks for all search terms
original_query = cleaned_keywords
search_links = []
for term in concept_mapping["search_terms"]:
search_link = f"/search?kw={term.replace(' ', '+')}&view=rich"
search_links.append(f"<a href='{search_link}'><strong>{term}</strong></a>")
links_html = ", ".join(search_links)
concept_notice = f"<div class='correction-notice'><strong>{original_query.capitalize()}</strong>: {concept_mapping['explanation']}<br>You may find what you are looking for with these search terms: {links_html}</div>"
# Use the first search term for the actual search
cleaned_keywords = concept_mapping["search_terms"][0]
def score_result_by_proximity(verse_text, query_words):
"""
Score a verse based on how close together the query words appear.
Higher score = words appear closer together.
"""
verse_lower = verse_text.lower()
# Find all positions of each query word
word_positions = {}
for word in query_words:
word_lower = word.lower()
positions = []
start = 0
while True:
pos = verse_lower.find(word_lower, start)
if pos == -1:
break
positions.append(pos)
start = pos + 1
word_positions[word] = positions
# If any word is missing, score is 0
if any(len(positions) == 0 for positions in word_positions.values()):
return 0
# Calculate minimum distance between all query words
# For each combination of positions, find the span (max_pos - min_pos)
# Smaller span = higher score
import itertools
# Get all combinations of positions for all words
all_position_combos = list(itertools.product(*word_positions.values()))
if not all_position_combos:
return 0
# Find the combination with minimum span
min_span = float('inf')
for combo in all_position_combos:
span = max(combo) - min(combo)
if span < min_span:
min_span = span
# Convert span to score (smaller span = higher score)
# Use inverse: 1000 / (span + 1) so adjacent words score ~1000, far apart score low
score = 1000.0 / (min_span + 1)
return score
results = lookup_fts(cleaned_keywords)
# Sort results by proximity score if we have multiple query words
if results and len(cleaned_keywords.split()) > 1:
query_words = cleaned_keywords.split()
scored_results = []
for result in results:
verse_text = result[4] # The text is at index 4
score = score_result_by_proximity(verse_text, query_words)
scored_results.append((score, result))
# Sort by score descending, then keep just the results
scored_results.sort(key=lambda x: x[0], reverse=True)
results = [result for score, result in scored_results]
if not results:
# PRIORITY 1: Try grammatical variant expansion first
words = cleaned_keywords.split()
@@ -904,9 +1011,45 @@ def search():
# PRIORITY 4: If we still have no results, show spelling suggestions
suggestions = h.suggest(cleaned_keywords)
return "{} not found, try one of the following: {}".format(
cleaned_keywords, ", ".join(suggestions) if suggestions else "No suggestions available"
)
arg_view = request.args.get("view", "rich")
if arg_view == "rich":
# Build a nice no-results page with header
no_results_html = cached["html_navbar"].format(
query=cleaned_keywords,
current_chapter_name="",
next_chapter_name="",
next_chapter_link="",
prev_chapter_name="",
prev_chapter_link=""
)
no_results_html += cached["css_navbar"]
no_results_html += "<div class='responses'>"
no_results_html += f"<div class='correction-notice'><strong>No results found</strong> for '{cleaned_keywords}'.</div>"
if suggestions:
no_results_html += "<div class='theological-suggestions'><h3>Did you mean?</h3>"
no_results_html += "<p>We couldn't find any verses matching your search. Here are some spelling suggestions:</p>"
for suggestion in suggestions[:5]: # Limit to 5 suggestions
search_link = f"/search?kw={suggestion}&view=rich"
no_results_html += f"<div class='suggestion-item'><a href='{search_link}'><strong>{suggestion}</strong></a></div>"
no_results_html += "</div>"
else:
no_results_html += "<div class='theological-suggestions'><h3>Search Tips</h3>"
no_results_html += "<p>We couldn't find any verses matching your search. Try:</p>"
no_results_html += "<ul style='font-family: var(--font-ui); color: var(--color-text-secondary);'>"
no_results_html += "<li>Checking your spelling</li>"
no_results_html += "<li>Using different keywords</li>"
no_results_html += "<li>Searching for <a href='/search?kw=love&view=rich'>love</a>, <a href='/search?kw=faith&view=rich'>faith</a>, or <a href='/search?kw=hope&view=rich'>hope</a></li>"
no_results_html += "</ul></div>"
no_results_html += "</div>"
return no_results_html
else:
# Plain text fallback
return "{} not found, try one of the following: {}".format(
cleaned_keywords, ", ".join(suggestions) if suggestions else "No suggestions available"
)
response_list = scripture_response(results)
# Check if any words in the search have theological synonyms to suggest
@@ -918,15 +1061,27 @@ def search():
for synonym in synonyms:
theological_suggestions.append(synonym)
arg_view = request.args.get("view", None)
arg_view = request.args.get("view", "rich")
if arg_view == "plain":
response = "<br>".join(response_list)
if concept_notice:
response = concept_notice + "<br>" + response
if theological_suggestions:
response += "<br><br>Also consider searching: "
for suggestion in theological_suggestions:
response += f"{suggestion}<br>"
elif arg_view == "rich":
response = "".join(response_list)
# Insert concept notice at the top if it exists
if concept_notice:
# Find the first </div> after the opening <div class='responses'> and insert after it
responses_start = response.find("<div class='responses'>")
if responses_start != -1:
# Find the results-count div closing
insert_pos = response.find("</div>", responses_start + len("<div class='responses'>"))
if insert_pos != -1:
insert_pos += len("</div>")
response = response[:insert_pos] + concept_notice + response[insert_pos:]
if theological_suggestions:
# Add suggestions before the LAST closing </div> which closes the responses container
suggestions_html = "<div class='theological-suggestions'><h3>Related Theological Concepts</h3>"