File size: 8,704 Bytes
d1ae506
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
"""
Get Results Rhode Island
Copyright (c) 2024 Cannlytics

Authors:
    Keegan Skeate <https://github.com/keeganskeate>
Created: 5/25/2024
Updated: 5/30/2024
License: CC-BY 4.0 <https://huggingface.co/datasets/cannlytics/cannabis_tests/blob/main/LICENSE>

Description:

    Curate Rhode Island lab result data obtained through public records requests.

Data Sources:
    
    - Public records request

"""
# Standard imports:
from datetime import datetime
import os

# External imports:
from cannlytics.data import save_with_copyright
from cannlytics.utils import snake_case
from cannlytics.utils.constants import ANALYTES
import numpy as np
import pandas as pd

# Define columns.
columns = {
    'Id': 'sample_id',
    'TestingFacilityName': 'lab',
    'ItemFromFacilityLicenseNumber': 'producer_license_number',
    'SourcePackageLabels': 'label',
    'TestPerformedDate': 'date_tested',
    'TestTypeName': 'test_type',
    'TestResultLevel': 'test_result',
    'OverallPassed': 'status',
}

# Define the data types for each column.
dtype_spec = {
    'Id': str,
    'TestingFacilityName': str,
    'ItemFromFacilityLicenseNumber': str,
    'SourcePackageLabels': str,
    'TestPerformedDate': str,
    'TestTypeName': str,
    'TestResultLevel': float,
    'OverallPassed': bool,
}

def collect_data(data_dir, columns, dtype_spec):
    """Collect data from a directory of CSV and Excel files."""
    results = []
    for root, _, files in os.walk(data_dir):
        for file in files:
            if 'no data' in file.lower():
                continue
            print('Reading:', file)
            file_path = os.path.join(root, file)
            if file.endswith('.csv'):
                df = read_and_standardize_csv(file_path, columns, dtype_spec)
            elif file.endswith('.xlsx'):
                df = read_and_standardize_excel(file_path, columns)
            if not df.empty:
                results.append(df)
    return pd.concat(results, ignore_index=True)

def read_and_standardize_csv(file_path, columns, dtype_spec):
    """Read a CSV file and standardize the column names."""
    try:
        df = pd.read_csv(file_path, dtype=dtype_spec, usecols=columns.keys(), encoding='latin1')
        df.rename(columns=columns, inplace=True)
        return df
    except Exception as e:
        print(f"Error reading {file_path}: {e}")
        return pd.DataFrame()

def read_and_standardize_excel(file_path, columns):
    """Read an Excel file and standardize the column names."""
    try:
        df = pd.read_excel(file_path, usecols=columns.keys())
        df.rename(columns=columns, inplace=True)
        return df
    except Exception as e:
        print(f"Error reading {file_path}: {e}")
        return pd.DataFrame()

def extract_test_details(data):
    """Extract test_name, units, and product_type from test_type."""
    data[['test_name', 'units', 'product_type']] = data['test_type'].str.extract(r'(.+?) \((.+?)\) (.+)')
    return data

def pivot_data(data):
    """Pivot the data to get results for each sample."""
    results = data.pivot_table(
        index=['sample_id', 'producer_license_number', 'lab', 'label', 'date_tested', 'product_type'],
        columns='test_name',
        values='test_result',
        aggfunc='first'
    ).reset_index()
    results['date_tested'] = pd.to_datetime(results['date_tested'], errors='coerce')
    results['month'] = results['date_tested'].dt.to_period('M')
    results['year'] = results['date_tested'].dt.year
    return results

def augment_calculations(
        df,
        cannabinoids=None,
        terpenes=None,
        delta_9_thc='delta_9_thc',
        thca='thca',
        cbd='cbd',
        cbda='cbda',
    ):
    """Augment the DataFrame with additional calculated fields."""
    # Calculate total cannabinoids.
    if cannabinoids is not None:
        df['total_cannabinoids'] = round(df[cannabinoids].sum(axis=1), 2)

    # Calculate total terpenes.
    if terpenes is not None:
        df['total_terpenes'] = round(df[terpenes].sum(axis=1), 2)

    # Calculate the total THC to total CBD ratio.
    df['total_thc'] = round(df[delta_9_thc] + 0.877 * df[thca], 2)
    df['total_cbd'] = round(df[cbd] + 0.877 * df[cbda], 2)
    df['thc_cbd_ratio'] = round(df['total_thc'] / df['total_cbd'], 2)

    # Calculate the total cannabinoids to total terpenes ratio.
    if cannabinoids is not None and terpenes is not None:
        df['cannabinoids_terpenes_ratio'] = round(df['total_cannabinoids'] / df['total_terpenes'], 2)

    # Return the augmented data.
    return df

def standardize_analyte_names(df, analyte_mapping):
    """Standardize analyte names."""
    df.columns = [analyte_mapping.get(snake_case(col), snake_case(col)) for col in df.columns]
    return df

def combine_similar_columns(df, similar_columns):
    """Combine similar columns with different spellings or capitalization."""
    for target_col, col_variants in similar_columns.items():
        if target_col not in df.columns:
            df[target_col] = pd.NA
        for col in col_variants:
            if col in df.columns:
                df[target_col] = df[target_col].combine_first(df[col])
                df.drop(columns=[col], inplace=True)
    return df

def get_results_ri(data_dir: str, output_dir: str) -> pd.DataFrame:

    # Collect Rhode Island lab results
    data = collect_data(data_dir, columns, dtype_spec)
    print('Number of Rhode Island tests:', len(data))

    # Extract test details
    data = extract_test_details(data)

    # Pivot the data to get results for each sample.
    results = pivot_data(data)
    print('Number of Rhode Island samples:', len(results))

    # Combine similar names.
    similar_columns = {
        'total_yeast_and_mold': ['Total Yeast and MOld', 'Total Yeast and Mold'],
        '1_2_dichloroethane': ['1,2 Dichlorethane', '1,2 Dichloroethane'],
        'total_cbd': ['Total CBD'],
        'total_thc': ['Total THC'],
        '3_methylpentane': ['3 Methylpetane', '3 Methylpentane'],
        'n_methylpyrrolidone': ['N Methylpyrrolidone', 'N methylpyrrlidone'],
        'n_n_dimethylacetamide': ['N,N Dimethyacetamide', 'N,N Dimethylacetamide'],
    }
    results = combine_similar_columns(results, similar_columns)

    # Standardize the analyte names
    results = standardize_analyte_names(results, ANALYTES)
    print('Standardized analyte names.')

    # Augment additional calculated metrics.
    cannabinoids = ['cbd', 'cbda', 'delta_9_thc', 'thca']
    terpenes = [
        'alpha_bisabolol', 'alpha_humulene', 'alpha_pinene',
        'alpha_terpinene', 'beta_caryophyllene', 'beta_myrcene',
        'beta_pinene', 'caryophyllene_oxide', 'd_limonene', 'linalool',
        'nerolidol', 'other_terpenes'
    ]
    results = augment_calculations(results)
    print('Augmented fields.')

    # Sort the columns.
    non_numeric = [
        'sample_id', 'producer_license_number', 'lab', 'label',
        'date_tested', 'product_type', 'month', 'year'
    ]
    numeric_cols = results.columns.difference(non_numeric)
    numeric_cols_sorted = sorted(numeric_cols)
    results = results[non_numeric + numeric_cols_sorted]

    # # Save the results with copyright and sources sheets.
    # date = datetime.now().strftime('%Y-%m-%d')
    # if not os.path.exists(output_dir): os.makedirs(output_dir)
    # outfile = f'{output_dir}/ri-results-{date}.xlsx'
    # save_with_copyright(
    #     results,
    #     outfile,
    #     dataset_name='Rhode Island Cannabis Lab Results',
    #     author='Keegan Skeate',
    #     publisher='Cannlytics',
    #     sources=['Rhode Island Office Of Cannabis Regulation'],
    #     source_urls=['https://dbr.ri.gov/office-cannabis-regulation'],
    # )
    # print('Saved Rhode Island lab results:', outfile)

    # Save the results.
    outfile = os.path.join(output_dir, 'ri-results-latest.xlsx')
    outfile_csv = os.path.join(output_dir, 'ri-results-latest.csv')
    outfile_json = os.path.join(output_dir, 'ri-results-latest.jsonl')
    results.to_excel(outfile, index=False)
    results.to_csv(outfile_csv, index=False)
    # FIXME: This causes an OverflowError
    # results.to_json(outfile_json, orient='records', lines=True)
    print('Saved Excel:', outfile)
    print('Saved CSV:', outfile_csv)
    # print('Saved JSON:', outfile_json)

    # Return the results.
    return results

# === Test ===
# [✓] Tested: 2024-07-10 by Keegan Skeate <keegan@cannlytics>
if __name__ == '__main__':

    # Define where the data lives.
    data_dir = 'D://data/public-records/Rhode Island/Rhode Island'
    output_dir = 'D://data/rhode-island/results/datasets'

    # Curate results.
    get_results_ri(data_dir=data_dir, output_dir=output_dir)