GotPsi / src /processors /cardS_processor.py
ebrinz's picture
Flatten history: re-sanitized (128-bit hash) GotPsi public dataset
9deebf2
Raw
History Blame Contribute Delete
13.9 kB
"""
CardS (Sequential Card Test) dataset processor.
Processes ESP sequential card finding test data with mixed row types:
- Step rows: 4 columns (user_id, trial, response, timestamp)
- Completion rows: 7 columns (user_id, trial, steps, response_array, target_image, timestamp)
"""
from pathlib import Path
from typing import List, Dict, Any, Optional
from datetime import datetime
import pandas as pd
import re
from ..core.base_classes import BaseProcessor
from ..core.config import Config
from ..core.exceptions import ProcessorError
from ..cleaners.encoding_cleaner import EncodingCleaner
from ..cleaners.delimiter_cleaner import DelimiterCleaner
from ..cleaners.temporal_parser import TemporalParser
class CardSProcessor(BaseProcessor):
"""
Processes CardS (Sequential Card Test) data.
Format: Mixed row types in same file
- Step rows: user clicks on a card position (1-5)
- Completion rows: trial finished, contains full response sequence
"""
BATCH_DELIMITER_CLEAN = False # mixed row formats (4 and 11 columns), no delimiter standardization
# File discovery patterns
FILE_PATTERNS = ['cardS*.dat']
# Column definitions for different row types
COLUMNS_STEP = [
'user_id', 'trial', 'response', 'timestamp'
]
COLUMNS_COMPLETION = [
'user_id', 'trial', 'steps', 'response_array', 'target_image', 'timestamp'
]
# Unified output schema
COLUMNS_UNIFIED = [
'user_id', 'trial', 'row_type',
'response', 'steps', 'response_array', 'target_image',
'timestamp', 'file_date', 'source_file', 'source_row_number'
]
def __init__(self, config: Config):
"""
Initialize CardS processor.
Args:
config: Configuration object
"""
super().__init__(config, 'cards')
# Initialize cleaners
self.encoding_cleaner = EncodingCleaner(config, self.errata_logger)
self.delimiter_cleaner = DelimiterCleaner(config, self.errata_logger)
self.temporal_parser = TemporalParser(config, self.errata_logger)
# Statistics
self.stats = {
'files_processed': 0,
'files_failed': 0,
'rows_total': 0,
'rows_step': 0,
'rows_completion': 0,
'rows_valid': 0,
'rows_invalid': 0
}
def extract_file_date(self, file_path: Path) -> Optional[datetime]:
"""
Extract date from filename (cardSYYMMDD.dat).
Args:
file_path: Path to file
Returns:
datetime object or None
"""
match = re.search(r'cardS(\d{6})\.dat', file_path.name)
if not match:
return None
date_str = match.group(1)
return self.temporal_parser.extract_date_from_filename(file_path.name)
def parse_trial_number(self, trial_str: str) -> int:
"""
Parse trial number, removing '.' suffix if present.
The '.' suffix indicates the first step of a trial.
Args:
trial_str: Trial string (e.g., "1.", "8.", "5")
Returns:
Integer trial number
"""
# Remove trailing '.' if present
trial_clean = str(trial_str).rstrip('.')
try:
return int(trial_clean)
except ValueError:
return None
def detect_row_type(self, row: pd.Series) -> str:
"""
Detect if row is a step or completion based on column count.
Args:
row: DataFrame row
Returns:
'step' or 'completion'
"""
# Count non-null columns
col_count = row.notna().sum()
if col_count >= 6: # Completion rows have 6-7 columns
return 'completion'
else: # Step rows have 4 columns
return 'step'
def parse_response_array(self, array_str: str) -> List[int]:
"""
Parse comma-separated response array.
Args:
array_str: String like "0,4,5,3,2,0"
Returns:
List of integers
"""
if pd.isna(array_str) or not array_str:
return []
try:
return [int(x.strip()) for x in str(array_str).split(',')]
except (ValueError, AttributeError):
return []
def process_file(self, file_path: Path, pre_cleaned_text: Optional[str] = None) -> Optional[pd.DataFrame]:
"""
Process a single CardS file.
Args:
file_path: Path to file
pre_cleaned_text: Pre-cleaned text from parallel batch cleaning.
If None, cleans the file inline (backward compat).
Returns:
DataFrame or None if processing fails
"""
try:
# Use pre-cleaned text if available, otherwise clean inline
if pre_cleaned_text is not None:
text = pre_cleaned_text
else:
text = self.encoding_cleaner.clean(file_path)
# DON'T use delimiter cleaner - CardS has mixed row formats (4 and 11 columns)
# which causes pandas to skip the 11-column rows as "bad lines"
# Parse manually to handle variable column counts (4 vs 11)
# Pandas read_csv skips rows with different column counts
from io import StringIO
import csv
reader = csv.reader(StringIO(text), skipinitialspace=True)
rows = list(reader)
if not rows:
self.errata_logger.log_error(
'empty_file',
'File is empty after CSV parsing - entire file omitted',
file_path=str(file_path),
scope='file'
)
return None
# Convert to DataFrame with max columns
df = pd.DataFrame(rows)
if df.empty:
self.errata_logger.log_error(
'empty_file',
'File is empty or has no valid rows - entire file omitted',
file_path=str(file_path),
scope='file'
)
return None
# Detect row types based on column count
# Step rows: user, trial, response, timestamp (4 columns)
# Completion rows: user, trial, steps, [array values 0-5], image, timestamp (11 columns)
# The response_array like "0,2,1,0,0,0" gets split into 6 separate columns
file_date = self.extract_file_date(file_path)
col_counts = df.notna().sum(axis=1)
is_step = col_counts == 4 # Step rows have exactly 4 columns
is_completion = col_counts >= 10 # Completion rows have 10-11 columns
# Count row types
self.stats['rows_step'] += is_step.sum()
self.stats['rows_completion'] += is_completion.sum()
# Create unified DataFrame
result_df = pd.DataFrame(index=df.index)
# Common columns for all rows
result_df['user_id'] = df.iloc[:, 0]
result_df['file_date'] = file_date
# Add audit columns (source file and row numbers)
result_df['source_file'] = file_path.name
result_df['source_row_number'] = range(1, len(result_df) + 1)
# Parse trial numbers (vectorized) - remove '.' suffix
result_df['trial'] = df.iloc[:, 1].astype(str).str.rstrip('.').astype(float)
# Process step rows (4 columns: user, trial, response, timestamp)
result_df.loc[is_step, 'row_type'] = 'step'
result_df.loc[is_step, 'response'] = df.loc[is_step].iloc[:, 2]
result_df.loc[is_step, 'steps'] = None
result_df.loc[is_step, 'response_array'] = None
result_df.loc[is_step, 'target_image'] = None
# Process completion rows (11 columns: user, trial, steps, arr0-arr5, image, timestamp)
# Recombine the 6 array values back into comma-separated string
result_df.loc[is_completion, 'row_type'] = 'completion'
result_df.loc[is_completion, 'response'] = None
result_df.loc[is_completion, 'steps'] = df.loc[is_completion].iloc[:, 2]
# Recombine columns 3-8 into response_array (vectorized)
if is_completion.any():
comp_df = df.loc[is_completion]
result_df.loc[is_completion, 'response_array'] = (
comp_df.iloc[:, 3].astype(str) + ',' +
comp_df.iloc[:, 4].astype(str) + ',' +
comp_df.iloc[:, 5].astype(str) + ',' +
comp_df.iloc[:, 6].astype(str) + ',' +
comp_df.iloc[:, 7].astype(str) + ',' +
comp_df.iloc[:, 8].astype(str)
)
result_df.loc[is_completion, 'target_image'] = comp_df.iloc[:, 9].values
# Parse timestamps: build a single raw series, parse once to avoid
# tz-aware dtype conflicts from partial .loc assignments
raw_ts = pd.Series(index=df.index, dtype='object')
if is_step.any():
raw_ts.loc[is_step] = df.loc[is_step].iloc[:, 3].values
if is_completion.any():
raw_ts.loc[is_completion] = df.loc[is_completion].iloc[:, 10].values
result_df['timestamp'] = self.temporal_parser.parse_dates_vectorized(
raw_ts, file_path=str(file_path)
)
# Validate and clean
result_df = self.validate_dataframe(result_df, file_path)
# Ensure all unified columns exist
for col in self.COLUMNS_UNIFIED:
if col not in result_df.columns:
result_df[col] = None
# Reorder to unified schema
result_df = result_df[self.COLUMNS_UNIFIED]
self.stats['files_processed'] += 1
self.stats['rows_valid'] += len(result_df)
return result_df
except Exception as e:
import traceback
self.errata_logger.log_error(
'file_processing_failed',
f'{type(e).__name__}: {e} - entire file omitted',
file_path=str(file_path),
scope='file',
context={'traceback': traceback.format_exc()},
)
return None
def validate_dataframe(self, df: pd.DataFrame, file_path: Path) -> pd.DataFrame:
"""
Validate and clean DataFrame.
Args:
df: Input DataFrame
file_path: Source file path
Returns:
Cleaned DataFrame
"""
initial_count = len(df)
valid_mask = pd.Series([True] * len(df), index=df.index)
# Validate user_id length
if 'user_id' in df.columns:
long_users = df['user_id'].str.len() > 64
if long_users.any():
valid_mask &= ~long_users
self.errata_logger.log_error(
'user_id_too_long',
f'{long_users.sum()} rows with user_id > 64 chars',
file_path=str(file_path),
scope='row'
)
# Validate response range (1-5) for step rows
if 'response' in df.columns:
df['response'] = self.to_numeric_logged(df['response'], 'response', str(file_path))
step_rows = df['row_type'] == 'step'
invalid_response = step_rows & ((df['response'] < 1) | (df['response'] > 5))
if invalid_response.any():
valid_mask &= ~invalid_response
self.errata_logger.log_error(
'response_out_of_range',
f'{invalid_response.sum()} rows with response not in 1-5',
file_path=str(file_path),
scope='row'
)
# Validate trial number (must be >= 1)
if 'trial' in df.columns:
df['trial'] = self.to_numeric_logged(df['trial'], 'trial', str(file_path))
invalid_trial = (df['trial'] < 1)
if invalid_trial.any():
valid_mask &= ~invalid_trial
self.errata_logger.log_error(
'trial_number_invalid',
f'{invalid_trial.sum()} rows with trial < 1',
file_path=str(file_path),
scope='row'
)
# Validate steps (1-5) for completion rows
if 'steps' in df.columns:
df['steps'] = self.to_numeric_logged(df['steps'], 'steps', str(file_path))
completion_rows = df['row_type'] == 'completion'
invalid_steps = completion_rows & ((df['steps'] < 1) | (df['steps'] > 5))
if invalid_steps.any():
valid_mask &= ~invalid_steps
self.errata_logger.log_error(
'steps_out_of_range',
f'{invalid_steps.sum()} rows with steps not in 1-5',
file_path=str(file_path),
scope='row'
)
# Convert numeric columns
numeric_cols = ['trial', 'response', 'steps']
for col in numeric_cols:
if col in df.columns:
df[col] = self.to_numeric_logged(df[col], col, str(file_path))
# Filter to valid rows
df_valid = df[valid_mask].copy()
invalid_count = initial_count - len(df_valid)
if invalid_count > 0:
self.errata_logger.log_file_summary(
str(file_path),
'partial',
initial_count,
len(df_valid),
invalid_count
)
return df_valid
def get_stats(self) -> Dict[str, Any]:
"""Get processing statistics."""
return self.stats.copy()