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
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499 | class ComorbidityDistributor:
"""High-performance distributor to assign comorbidities to people based on their age, sex, ethnicity,
and region. Uses vectorized operations and batch processing for UK-scale simulations.
"""
def __init__(self):
"""
Initialize the OptimizedComorbidityDistributor with data path and configuration.
"""
self.script_dir = Path(__file__).parent.parent.parent # Go up to june-measles root
self.data_path = self.script_dir / "data" / "input" / "demography" / "comorbidities" / "processed_counts_midpoint_rounded_proportions_n.csv"
# Optimized data structures
self._comorbidity_data = None
self._data_loaded = False
self._demographics_cache = {}
self._probability_arrays = {}
# Define condition columns from the CSV
self.condition_columns = [
'has_had_cvd_diagnosis_count_midpoint_rounded',
'has_had_crd_diagnosis_count_midpoint_rounded',
'has_had_ckd_diagnosis_count_midpoint_rounded',
'has_had_cld_diagnosis_count_midpoint_rounded',
'severe_obesity_count_midpoint_rounded',
'has_had_cancer_diagnosis_count_midpoint_rounded',
'has_had_immunosuppression_diagnosis_count_midpoint_rounded',
'has_had_neuro_diagnosis_count_midpoint_rounded'
]
# Map condition columns to readable names
self.condition_names = {
'has_had_cvd_diagnosis_count_midpoint_rounded': 'cardiovascular_disease',
'has_had_crd_diagnosis_count_midpoint_rounded': 'chronic_respiratory_disease',
'has_had_ckd_diagnosis_count_midpoint_rounded': 'chronic_kidney_disease',
'has_had_cld_diagnosis_count_midpoint_rounded': 'chronic_liver_disease',
'severe_obesity_count_midpoint_rounded': 'severe_obesity',
'has_had_cancer_diagnosis_count_midpoint_rounded': 'cancer',
'has_had_immunosuppression_diagnosis_count_midpoint_rounded': 'immunosuppression',
'has_had_neuro_diagnosis_count_midpoint_rounded': 'neurological_condition'
}
def _load_comorbidity_data(self):
"""Load and optimize the comorbidity prevalence data from CSV file."""
if self._data_loaded:
return
logger.info("Loading and optimizing comorbidity prevalence data...")
try:
if not self.data_path.exists():
raise FileNotFoundError(f"Comorbidity data file not found at {self.data_path}")
# Load CSV with optimized dtypes
dtype_dict = {
'sex': 'category',
'age_band_min': 'category',
'combined_ethnicity_less': 'category',
'region': 'category'
}
self._comorbidity_data = pd.read_csv(self.data_path, dtype=dtype_dict)
logger.info(f"Loaded comorbidity data: {len(self._comorbidity_data)} demographic groups")
# Create optimized multi-index lookup
self._comorbidity_data.set_index(['sex', 'age_band_min', 'combined_ethnicity_less', 'region'],
inplace=True)
# Pre-compute probability matrices for vectorized operations
self._precompute_probability_matrices()
self._data_loaded = True
logger.info("Comorbidity data optimization completed")
except Exception as e:
logger.error(f"Error loading comorbidity data: {e}")
self._data_loaded = False
raise
def _precompute_probability_matrices(self):
"""Pre-compute probability matrices for faster vectorized lookups."""
logger.info("Pre-computing probability matrices...")
# Create fast lookup arrays for each demographic combination
for idx, row in self._comorbidity_data.iterrows():
key = idx # Multi-index tuple
# Store probabilities as numpy array for fast access
probs = np.array([row[col] for col in self.condition_columns])
has_any_prob = row.get('has_comorbidity_midpoint_rounded', 0.0)
multiple_prob = row.get('multiple_morbidities_count_midpoint_rounded', 0.0)
self._probability_arrays[key] = {
'condition_probs': probs,
'has_any_prob': has_any_prob,
'multiple_prob': multiple_prob,
'condition_names': list(self.condition_names.values())
}
def _get_person_demographics(self, person: Person) -> Tuple[str, str, str, str]:
"""Extract and cache person demographics in optimized format.
Args:
person (Person): The person object
Returns:
Tuple[str, str, str, str]: (sex, age_band, ethnicity, region) tuple
"""
# Convert sex to CSV format
sex = "female" if person.sex == "f" else "male"
# Get age band
age = person.age
if age < 10:
age_band = "0-9"
elif age < 18:
age_band = "10-17"
elif age < 30:
age_band = "18-29"
elif age < 50:
age_band = "30-49"
elif age < 75:
age_band = "50-74"
else:
age_band = "75-99"
# Get ethnicity and map to CSV format
ethnicity = getattr(person, 'ethnicity', 'W')
if ethnicity == 'O': # Other -> Combined Other
ethnicity = 'CO'
# Get region with mapping
region = self._get_person_region_fast(person)
return sex, age_band, ethnicity, region
def _get_person_region_fast(self, person: Person) -> str:
"""Fast region extraction with caching.
Args:
person (Person): The person object
Returns:
str: The region name with mappings applied
"""
# Use caching for region lookups
person_id = getattr(person, 'id', None)
if person_id and person_id in self._demographics_cache:
return self._demographics_cache[person_id]['region']
region_name = None
# Try to get region from person's household or area
if hasattr(person, 'household') and person.household:
if hasattr(person.household, 'area') and person.household.area:
if hasattr(person.household.area, 'region'):
region_name = person.household.area.region.name
# Try alternative attributes
if not region_name and hasattr(person, 'area') and person.area:
if hasattr(person.area, 'region'):
region_name = person.area.region.name
# Apply regional mappings
if region_name:
region_lower = region_name.lower()
# Map non-English regions to English regions
if region_lower == 'scotland':
region_name = 'North East'
elif region_lower == 'wales':
region_name = 'North West'
elif region_lower in ['northern ireland', 'northernireland']:
region_name = 'North West'
else:
# Map full UK region names to CSV shorthand
region_mappings = {
'east of england': 'East',
'east midlands': 'East Midlands',
'greater london': 'London',
'london': 'London',
'north east': 'North East',
'north west': 'North West',
'south east': 'South East',
'south west': 'South West',
'west midlands': 'West Midlands',
'yorkshire and the humber': 'Yorkshire and The Humber',
'yorkshire and humber': 'Yorkshire and The Humber'
}
mapped_region = region_mappings.get(region_lower)
if mapped_region:
region_name = mapped_region
return region_name or 'London' # Default fallback
def _group_people_by_demographics(self, people: List[Person]) -> Dict[Tuple, List[Person]]:
"""Group people by their demographic characteristics for batch processing.
Args:
people (List[Person]): List of people to group
Returns:
Dict[Tuple, List[Person]]: Dictionary mapping demographic tuples to lists of people
"""
groups = defaultdict(list)
for person in people:
demographics = self._get_person_demographics(person)
groups[demographics].append(person)
return dict(groups)
def _assign_comorbidities_to_group(self, people: List[Person], demographics: Tuple) -> None:
"""Assign comorbidities to a group of people with same demographics using vectorized operations.
Args:
people (List[Person]): List of people with same demographics
demographics (Tuple): (sex, age_band, ethnicity, region) tuple
"""
if not people:
return
n_people = len(people)
sex, age_band, ethnicity, region = demographics
# Try exact match first
if demographics in self._probability_arrays:
prob_data = self._probability_arrays[demographics]
else:
# Find fallback data efficiently
prob_data = self._find_fallback_data_optimized(demographics)
if not prob_data:
# Use age-based defaults
prob_data = self._get_default_probabilities_optimized(age_band)
# Vectorized probability calculations
condition_probs = prob_data['condition_probs']
has_any_prob = prob_data['has_any_prob']
multiple_prob = prob_data['multiple_prob']
condition_names = prob_data['condition_names']
# Vectorized random sampling
has_any_conditions = np.random.random(n_people) < has_any_prob
has_multiple = np.random.random(n_people) < multiple_prob
# For people who should have conditions, sample which ones
for i, person in enumerate(people):
conditions = set()
if has_any_conditions[i]:
# Sample individual conditions
condition_mask = np.random.random(len(condition_probs)) < condition_probs
candidate_conditions = [condition_names[j] for j, has_condition in enumerate(condition_mask) if has_condition]
if candidate_conditions:
if has_multiple[i] and len(candidate_conditions) > 1:
# Keep multiple conditions
conditions.update(candidate_conditions)
else:
# Keep only one condition
conditions.add(np.random.choice(candidate_conditions))
elif has_multiple[i]:
# Force assign at least one condition if they should have multiple
# Choose based on probabilities
if condition_probs.sum() > 0:
normalized_probs = condition_probs / condition_probs.sum()
chosen_idx = np.random.choice(len(condition_names), p=normalized_probs)
conditions.add(condition_names[chosen_idx])
# Store conditions
person.comorbidity = conditions
def _find_fallback_data_optimized(self, demographics: Tuple) -> Optional[Dict]:
"""Optimized fallback data lookup using pandas operations.
Args:
demographics (Tuple): (sex, age_band, ethnicity, region) tuple
Returns:
Dict or None: Fallback probability data
"""
sex, age_band, ethnicity, region = demographics
# Try different fallback strategies efficiently using pandas indexing
fallback_strategies = [
# Same demographics, different region
[(sex, age_band, ethnicity), 3], # Don't filter region
# Same sex/age, White ethnicity, same region
[(sex, age_band, 'W', region), None],
# Same sex/age, White ethnicity, any region
[(sex, age_band, 'W'), 3],
# Same age, any sex/ethnicity, same region
[(None, age_band, None, region), [0, 2]], # Don't filter sex and ethnicity
# Same age, any sex/ethnicity/region
[(None, age_band, None), [0, 2, 3]] # Only filter age
]
for strategy_tuple, skip_levels in fallback_strategies:
try:
if skip_levels is None:
# Exact match attempt
if strategy_tuple in self._comorbidity_data.index:
row = self._comorbidity_data.loc[strategy_tuple].iloc[0] if isinstance(
self._comorbidity_data.loc[strategy_tuple], pd.DataFrame) else self._comorbidity_data.loc[strategy_tuple]
return self._row_to_prob_data(row)
else:
# Partial match - find any matching record
data_subset = self._comorbidity_data
for i, val in enumerate(strategy_tuple):
if val is not None and i not in (skip_levels if isinstance(skip_levels, list) else [skip_levels]):
level_name = data_subset.index.names[i]
data_subset = data_subset[data_subset.index.get_level_values(level_name) == val]
if not data_subset.empty:
row = data_subset.iloc[0]
logger.info(f"Using fallback data for {demographics}")
return self._row_to_prob_data(row)
except (KeyError, IndexError):
continue
return None
def _row_to_prob_data(self, row) -> Dict:
"""Convert DataFrame row to probability data format.
Args:
row (pandas.Series): DataFrame row
Returns:
Dict: Probability data dictionary
"""
condition_probs = np.array([row[col] for col in self.condition_columns])
has_any_prob = row.get('has_comorbidity_midpoint_rounded', 0.0)
multiple_prob = row.get('multiple_morbidities_count_midpoint_rounded', 0.0)
return {
'condition_probs': condition_probs,
'has_any_prob': has_any_prob,
'multiple_prob': multiple_prob,
'condition_names': list(self.condition_names.values())
}
def _get_default_probabilities_optimized(self, age_band: str) -> Dict:
"""Get optimized default comorbidity probabilities when no data is available.
Args:
age_band (str): Age band string
Returns:
Dict: Default probability data
"""
# Extract numeric age from age band
age = int(age_band.split('-')[0])
# Age-based default probabilities
if age < 18:
base_prob = 0.01
elif age < 30:
base_prob = 0.05
elif age < 50:
base_prob = 0.15
elif age < 75:
base_prob = 0.30
else:
base_prob = 0.50
condition_probs = np.full(len(self.condition_columns), base_prob * 0.1)
return {
'condition_probs': condition_probs,
'has_any_prob': base_prob,
'multiple_prob': base_prob * 0.1,
'condition_names': list(self.condition_names.values())
}
def assign_comorbidities_to_all_residents(self, world) -> None:
"""Optimized assignment of comorbidities to all residents in the world using batch processing.
Args:
world (World): The world object containing all groups
"""
logger.info("Starting comorbidity assignment for all residents...")
# Load data if not already loaded
if not self._data_loaded:
self._load_comorbidity_data()
total_people_processed = 0
condition_counts = {name: 0 for name in self.condition_names.values()}
multiple_conditions_count = 0
# Collect all people for batch processing
all_people = []
# Process households
if world.households is not None:
logger.info("Collecting household residents...")
for household in world.households:
if household.residents:
all_people.extend(household.residents)
# Process care homes
if world.care_homes is not None:
logger.info("Collecting care home residents...")
for care_home in world.care_homes:
if hasattr(care_home, 'residents') and care_home.residents:
all_people.extend(care_home.residents)
# Process boarding schools
if hasattr(world, 'boarding_schools') and world.boarding_schools is not None:
logger.info("Collecting boarding school residents...")
for boarding_school in world.boarding_schools:
if hasattr(boarding_school, 'residents') and boarding_school.residents:
all_people.extend(boarding_school.residents)
# Process student dorms
if world.student_dorms is not None:
logger.info("Collecting student dorm residents...")
for student_dorm in world.student_dorms:
if hasattr(student_dorm, 'residents') and student_dorm.residents:
all_people.extend(student_dorm.residents)
logger.info(f"Collected {len(all_people):,} people for batch processing")
# Group people by demographics for batch processing
logger.info("Grouping people by demographics...")
demographic_groups = self._group_people_by_demographics(all_people)
logger.info(f"Created {len(demographic_groups):,} demographic groups")
# Process each demographic group in batch
for i, (demographics, people_group) in enumerate(demographic_groups.items()):
if i % 100 == 0 and i > 0:
logger.info(f"Processed {i:,}/{len(demographic_groups):,} demographic groups")
self._assign_comorbidities_to_group(people_group, demographics)
total_people_processed += len(people_group)
# Update statistics
for person in people_group:
conditions = getattr(person, 'comorbidity', set())
for condition in conditions:
if condition in condition_counts:
condition_counts[condition] += 1
if len(conditions) > 1:
multiple_conditions_count += 1
logger.info(f"Completed optimized comorbidity assignment: {total_people_processed:,} people processed")
# Display statistics
self._display_assignment_statistics(total_people_processed, condition_counts, multiple_conditions_count)
def _display_assignment_statistics(self, total_people: int, condition_counts: Dict[str, int], multiple_count: int) -> None:
"""Display statistics about comorbidity assignments.
Args:
total_people (int): Total number of people processed
condition_counts (Dict[str, int]): Count of each condition assigned
multiple_count (int): Number of people with multiple conditions
"""
print("\n===== Comorbidity Assignment Statistics =====")
print(f"Total people processed: {total_people:,}")
people_with_any_condition = len([count for count in condition_counts.values() if count > 0])
total_conditions_assigned = sum(condition_counts.values())
print(f"People with any comorbidity: {total_conditions_assigned - multiple_count:,} ({(total_conditions_assigned - multiple_count)/total_people*100:.1f}%)")
print(f"People with multiple comorbidities: {multiple_count:,} ({multiple_count/total_people*100:.1f}%)")
print("\nCondition prevalence:")
for condition, count in sorted(condition_counts.items()):
if count > 0:
percentage = count / total_people * 100
print(f" {condition.replace('_', ' ').title()}: {count:,} ({percentage:.2f}%)")
print("="*60)
|