Stage 0: Data Extraction from SQLite

Primary author: Victoria Based on: George Ho's dataset schema (cryptics.georgeho.org) Prompt engineering: Victoria AI assistance: Claude (Anthropic) Environment: Local or Colab

This notebook extracts six tables from the raw data.sqlite3 database (downloaded from cryptics.georgeho.org) and saves them as CSV files with consistent, descriptive column names. These CSVs are the starting point for all downstream notebooks.

Dataset created by George Ho.

Available for download: https://cryptics.georgeho.org/data.db

This notebook takes the complete database downloaded from cryptics.georgeho.org, and outputs the six most relevant tables as .csv files.

It changes the feature names so that they are descriptive and consistent across all tables:

• rowid becomes clue_id, ind_id, or charade_id
• answer becomes charade_answer when it refers to the charade, remains answer when referring to the clue answer
• clue_rowids becomes clue_ids for consistency

# imports
import os
import sqlite3
import pandas as pd
import numpy as np
from pathlib import Path

# --- Environment Auto-Detection ---
try:
    IS_COLAB = 'google.colab' in str(get_ipython())
except NameError:
    IS_COLAB = False

if IS_COLAB:
    from google.colab import drive
    drive.mount('/content/drive')
    PROJECT_ROOT = Path('/content/drive/MyDrive/SIADS 692 Milestone II/Milestone II - NLP Cryptic Crossword Clues')
else:
    try:
        PROJECT_ROOT = Path(__file__).resolve().parent.parent
    except NameError:
        PROJECT_ROOT = Path.cwd().parent

DATA_DIR = PROJECT_ROOT / "data"

print(f'Project root: {PROJECT_ROOT}')
print(f'Data directory: {DATA_DIR}')

Project root: /home/vwinters/ccc-project/indicator_clustering
Data directory: /home/vwinters/ccc-project/indicator_clustering/data

# Connect to the sqlite3 file
data_file = str(DATA_DIR / "data.sqlite3")
conn = sqlite3.connect(data_file)

# Uncomment to see what data tables exist in the file
tables = pd.read_sql("SELECT name FROM sqlite_master WHERE type='table';", conn)
#tables

# Keep track of all tables that might be of interest from the original dataset
# Display the names and sizes of all tables.

tables = [
    "clues",
    "indicators",
    "charades",
    "indicators_by_clue",
    "charades_by_clue",
    "indicators_consolidated"
]

summary = []

for t in tables:
    # count rows
    row_count = pd.read_sql(f"SELECT COUNT(*) AS n FROM {t};", conn).iloc[0]["n"]
    
    # count rows and columns
    col_info = pd.read_sql(f"PRAGMA table_info({t});", conn)
    col_count = len(col_info)

    summary.append({
        "table": t,
        "rows": row_count,
        "columns": col_count
    })

summary_df = pd.DataFrame(summary)
summary_df.style.format({"rows": "{:,}"}) # display with commas 

	table	rows	columns
0	clues	660,613	9
1	indicators	15,735	4
2	charades	57,289	4
3	indicators_by_clue	88,037	9
4	charades_by_clue	126,825	3
5	indicators_consolidated	1	8

# Create the dataframes related to indicators
df_indicators = pd.read_sql("SELECT * FROM indicators;", conn)
df_ind_by_clue = pd.read_sql("SELECT * FROM indicators_by_clue;", conn)
df_indicators_consolidated = pd.read_sql("SELECT * FROM indicators_consolidated;", conn)

# Create dataframes pertaining to clue and charade
df_clues = pd.read_sql("SELECT * FROM clues;", conn)
df_charades = pd.read_sql("SELECT * FROM charades;", conn)
df_charades_by_clue = pd.read_sql("SELECT * FROM charades_by_clue;", conn)

# Indicators
display(df_indicators.head(3))

# Rename columns for consistency across tables/dataframes
df_indicators = df_indicators.rename(columns={'rowid': 'ind_id', 'clue_rowids': 'clue_ids'})
display(df_indicators.head(3))

	rowid	wordplay	indicator	clue_rowids
0	1	alternation	abnormal	[623961](/data/clues/623961)
1	2	alternation	after odd losses	[139327](/data/clues/139327)
2	3	alternation	after regular excisions	[107211](/data/clues/107211)

	ind_id	wordplay	indicator	clue_ids
0	1	alternation	abnormal	[623961](/data/clues/623961)
1	2	alternation	after odd losses	[139327](/data/clues/139327)
2	3	alternation	after regular excisions	[107211](/data/clues/107211)

# Indicators by Clue
display(df_ind_by_clue.head(4))
print()

# Rename columns for consistancy across tables/dataframes
df_ind_by_clue = df_ind_by_clue.rename(columns={'clue_rowid': 'clue_id'})
display(df_ind_by_clue.head(4))

# See how many contextualized indicators are in this table
#print("Instances of each CONTEXTUALIZED wordplay (multiple per clue, redundant indicators)")
#df_ind_by_clue.replace("", np.nan).count()

	clue_rowid	alternation	anagram	container	deletion	hidden	homophone	insertion	reversal
0	90		transforming
1	97		ground
2	101						verbally
3	142								about to go back

	clue_id	alternation	anagram	container	deletion	hidden	homophone	insertion	reversal
0	90		transforming
1	97		ground
2	101						verbally
3	142								about to go back

# Indicators Consolidated
display(df_indicators_consolidated.head())
print()

	alternation	anagram	container	deletion	hidden	homophone	insertion	reversal
0	abnormal\nafter odd losses\nafter regular exci...	a bad way\na bit differently\na brew of\na coc...	a single\naboard\nabout\nabout t\nabout/confin...	a certain amount off\nabandoned\nabandoned by\...	a bit\na bit of\na bit of this\na bunch of\na ...	a report on\naccording to announcement\naccord...	a bit of\na contingent of\na feature of\na fic...	a cleric raised\na fastening device put up\na ...

# Clues
display(df_clues.head(3))

# Rename columns for consistency across tables/dataframes
df_clues = df_clues.rename(columns={'rowid': 'clue_id'})
display(df_clues.head(3))

	rowid	clue	answer	definition	clue_number	puzzle_date	puzzle_name	source_url	source
0	1	Acquisitive chap, as we see it (8)	COVETOUS	Acquisitive	1a	2019-08-08	Times 27424	https://times-xwd-times.livejournal.com/218581...	times_xwd_times
1	2	Back yard fencing weak and sagging (6)	DROOPY	sagging	5a	2019-08-08	Times 27424	https://times-xwd-times.livejournal.com/218581...	times_xwd_times
2	3	Stripping off uniform, love holding colonel's ...	UNCLOTHING	Stripping	8a	2019-08-08	Times 27424	https://times-xwd-times.livejournal.com/218581...	times_xwd_times

	clue_id	clue	answer	definition	clue_number	puzzle_date	puzzle_name	source_url	source
0	1	Acquisitive chap, as we see it (8)	COVETOUS	Acquisitive	1a	2019-08-08	Times 27424	https://times-xwd-times.livejournal.com/218581...	times_xwd_times
1	2	Back yard fencing weak and sagging (6)	DROOPY	sagging	5a	2019-08-08	Times 27424	https://times-xwd-times.livejournal.com/218581...	times_xwd_times
2	3	Stripping off uniform, love holding colonel's ...	UNCLOTHING	Stripping	8a	2019-08-08	Times 27424	https://times-xwd-times.livejournal.com/218581...	times_xwd_times

# Charades
display(df_charades.head(3))

# Rename columns for consistency across tables/dataframes
df_charades = df_charades.rename(columns={'rowid': 'charade_id', 'answer':'charade_answer', 'clue_rowids': 'clue_ids'})
display(df_charades.head(3))

	rowid	charade	answer	clue_rowids
0	1	a	ALPHA	[108948](/data/clues/108948)
1	2	a	AN	[79856](/data/clues/79856), [162234](/data/clu...
2	3	a	AR	[384876](/data/clues/384876)

	charade_id	charade	charade_answer	clue_ids
0	1	a	ALPHA	[108948](/data/clues/108948)
1	2	a	AN	[79856](/data/clues/79856), [162234](/data/clu...
2	3	a	AR	[384876](/data/clues/384876)

# Charades by Clue
display(df_charades_by_clue.head(3))

# Rename columns for consistency across tables/dataframes
df_charades_by_clue = df_charades_by_clue.rename(columns={'clue_rowid': 'clue_id', 'answer':'charade_answer'})
display(df_charades_by_clue.head(3))

	clue_rowid	charade	answer
0	1	chap	COVE
1	1	as we see it	TO US
2	21	give the thing acclaim	LAUD IT

	clue_id	charade	charade_answer
0	1	chap	COVE
1	1	as we see it	TO US
2	21	give the thing acclaim	LAUD IT

# Write each dataframe to a CSV file in the data directory (without the index)
df_indicators.to_csv(DATA_DIR / "indicators_raw.csv", index=False)
df_ind_by_clue.to_csv(DATA_DIR / "indicators_by_clue_raw.csv", index=False)
df_indicators_consolidated.to_csv(DATA_DIR / "indicators_consolidated_raw.csv", index=False)
df_clues.to_csv(DATA_DIR / "clues_raw.csv", index=False)
df_charades.to_csv(DATA_DIR / "charades_raw.csv", index=False)
df_charades_by_clue.to_csv(DATA_DIR / "charades_by_clue_raw.csv", index=False)