Update hedis_engine.py
Browse files- hedis_engine.py +276 -23
hedis_engine.py
CHANGED
|
@@ -1,18 +1,16 @@
|
|
| 1 |
# hedis_engine.py
|
|
|
|
| 2 |
import os
|
|
|
|
| 3 |
from crewai import Agent, Task, Crew, Process
|
| 4 |
from crewai_tools import tool
|
| 5 |
from langchain_community.vectorstores import Chroma
|
| 6 |
-
|
| 7 |
from langchain_openai import ChatOpenAI
|
| 8 |
|
| 9 |
|
| 10 |
-
|
| 11 |
-
|
| 12 |
-
|
| 13 |
-
class HedisComplianceEngine:
|
| 14 |
"""
|
| 15 |
-
Crew that: (1) states official HEDIS measure criteria,
|
| 16 |
(2) extracts patient evidence from embeddings, and
|
| 17 |
(3) produces a markdown analysis report.
|
| 18 |
"""
|
|
@@ -21,6 +19,7 @@ class HedisComplianceEngine:
|
|
| 21 |
self.measure_code = measure_code
|
| 22 |
self.measurement_year = measurement_year
|
| 23 |
self.model = ChatOpenAI(model="gpt-4o", temperature=0, max_tokens=5096)
|
|
|
|
| 24 |
|
| 25 |
@tool("patient_chart_search")
|
| 26 |
def patient_chart_search(query: str) -> str:
|
|
@@ -70,16 +69,248 @@ class HedisComplianceEngine:
|
|
| 70 |
llm=self.model,
|
| 71 |
)
|
| 72 |
|
| 73 |
-
|
| 74 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 75 |
description=(
|
| 76 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 77 |
"- Required tests or procedures\n"
|
| 78 |
"- Diagnosis/procedure codes\n"
|
| 79 |
"- Required medications\n"
|
| 80 |
"- Inclusion and exclusion criteria\n"
|
| 81 |
-
"- Timeframes for each requirement (e.g., within
|
| 82 |
-
"You MUST compute and include absolute date ranges using {measurement_year}.\n"
|
| 83 |
"Return JSON with fields: 'measure','required_tests','required_medications',"
|
| 84 |
"'codes','timeframes','inclusions','exclusions'."
|
| 85 |
),
|
|
@@ -88,20 +319,20 @@ class HedisComplianceEngine:
|
|
| 88 |
async_execution=True,
|
| 89 |
)
|
| 90 |
|
| 91 |
-
|
| 92 |
description=(
|
| 93 |
-
"Using the patient_chart_search tool, extract only the tests, diagnoses, procedures, "
|
| 94 |
-
"medications, and dates relevant to {measure_code}. Only include measure related info from the patient charts"
|
| 95 |
),
|
| 96 |
expected_output="Structured list of patient evidence with dates.",
|
| 97 |
agent=self.patient_analyzer,
|
| 98 |
async_execution=True,
|
| 99 |
)
|
| 100 |
|
| 101 |
-
|
| 102 |
description=(
|
| 103 |
-
"Create a comprehensive markdown report analyzing the patient chart for HEDIS {measure_code} "
|
| 104 |
-
"for measurement year {measurement_year}.\n\n"
|
| 105 |
"Sections:\n"
|
| 106 |
"### Diagnoses Found\n"
|
| 107 |
"### Procedures Found\n"
|
|
@@ -115,20 +346,42 @@ class HedisComplianceEngine:
|
|
| 115 |
),
|
| 116 |
expected_output="A thorough markdown report of HEDIS findings.",
|
| 117 |
agent=self.hedis_chart_reporter,
|
| 118 |
-
context=[
|
| 119 |
)
|
| 120 |
|
| 121 |
-
|
| 122 |
agents=[self.hedis_expert, self.patient_analyzer, self.hedis_chart_reporter],
|
| 123 |
-
tasks=[
|
| 124 |
process=Process.sequential,
|
| 125 |
verbose=True,
|
| 126 |
)
|
| 127 |
|
| 128 |
-
def run(self) -> str:
|
| 129 |
inputs = {
|
| 130 |
"measure_code": self.measure_code,
|
| 131 |
"measurement_year": self.measurement_year,
|
| 132 |
}
|
| 133 |
-
result =
|
| 134 |
-
return str(result)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
# hedis_engine.py
|
| 2 |
+
# interactive_hedis_engine.py
|
| 3 |
import os
|
| 4 |
+
import json
|
| 5 |
from crewai import Agent, Task, Crew, Process
|
| 6 |
from crewai_tools import tool
|
| 7 |
from langchain_community.vectorstores import Chroma
|
|
|
|
| 8 |
from langchain_openai import ChatOpenAI
|
| 9 |
|
| 10 |
|
| 11 |
+
class InteractiveHedisComplianceEngine:
|
|
|
|
|
|
|
|
|
|
| 12 |
"""
|
| 13 |
+
Interactive Crew that: (1) states official HEDIS measure criteria and gets user feedback,
|
| 14 |
(2) extracts patient evidence from embeddings, and
|
| 15 |
(3) produces a markdown analysis report.
|
| 16 |
"""
|
|
|
|
| 19 |
self.measure_code = measure_code
|
| 20 |
self.measurement_year = measurement_year
|
| 21 |
self.model = ChatOpenAI(model="gpt-4o", temperature=0, max_tokens=5096)
|
| 22 |
+
self.hedis_criteria = None # Store approved criteria
|
| 23 |
|
| 24 |
@tool("patient_chart_search")
|
| 25 |
def patient_chart_search(query: str) -> str:
|
|
|
|
| 69 |
llm=self.model,
|
| 70 |
)
|
| 71 |
|
| 72 |
+
def get_hedis_criteria_with_feedback(self, suggestions: str = "") -> dict:
|
| 73 |
+
"""
|
| 74 |
+
Get HEDIS criteria from expert and present to user for feedback.
|
| 75 |
+
Returns approved criteria or None if user exits.
|
| 76 |
+
"""
|
| 77 |
+
# Create task for getting HEDIS criteria
|
| 78 |
+
hedis_task_description = (
|
| 79 |
+
f"Provide the official criteria for HEDIS measure {self.measure_code}, including:\n"
|
| 80 |
+
"- Required tests or procedures\n"
|
| 81 |
+
"- Diagnosis/procedure codes\n"
|
| 82 |
+
"- Required medications\n"
|
| 83 |
+
"- Inclusion and exclusion criteria\n"
|
| 84 |
+
f"- Timeframes for each requirement (e.g., within X years of {self.measurement_year})\n\n"
|
| 85 |
+
f"You MUST compute and include absolute date ranges using {self.measurement_year}.\n"
|
| 86 |
+
)
|
| 87 |
+
|
| 88 |
+
if suggestions:
|
| 89 |
+
hedis_task_description += f"\n\nIncorporate these suggestions: {suggestions}\n"
|
| 90 |
+
|
| 91 |
+
hedis_task_description += (
|
| 92 |
+
"Return JSON with fields: 'measure','required_tests','required_medications',"
|
| 93 |
+
"'codes','timeframes','inclusions','exclusions'."
|
| 94 |
+
)
|
| 95 |
+
|
| 96 |
+
hedis_task = Task(
|
| 97 |
+
description=hedis_task_description,
|
| 98 |
+
expected_output="Structured JSON with absolute date ranges.",
|
| 99 |
+
agent=self.hedis_expert,
|
| 100 |
+
)
|
| 101 |
+
|
| 102 |
+
# Create temporary crew just for getting criteria
|
| 103 |
+
criteria_crew = Crew(
|
| 104 |
+
agents=[self.hedis_expert],
|
| 105 |
+
tasks=[hedis_task],
|
| 106 |
+
process=Process.sequential,
|
| 107 |
+
verbose=True,
|
| 108 |
+
)
|
| 109 |
+
|
| 110 |
+
# Execute to get criteria
|
| 111 |
+
inputs = {
|
| 112 |
+
"measure_code": self.measure_code,
|
| 113 |
+
"measurement_year": self.measurement_year,
|
| 114 |
+
}
|
| 115 |
+
|
| 116 |
+
result = criteria_crew.kickoff(inputs=inputs)
|
| 117 |
+
|
| 118 |
+
try:
|
| 119 |
+
# Try to parse as JSON
|
| 120 |
+
criteria_json = json.loads(str(result))
|
| 121 |
+
return criteria_json
|
| 122 |
+
except json.JSONDecodeError:
|
| 123 |
+
# If not valid JSON, return as string
|
| 124 |
+
return {"raw_output": str(result)}
|
| 125 |
+
|
| 126 |
+
def display_criteria(self, criteria: dict) -> None:
|
| 127 |
+
"""Display the HEDIS criteria in a user-friendly format."""
|
| 128 |
+
print(f"\n{'='*60}")
|
| 129 |
+
print(f"HEDIS MEASURE {self.measure_code} CRITERIA")
|
| 130 |
+
print(f"Measurement Year: {self.measurement_year}")
|
| 131 |
+
print(f"{'='*60}")
|
| 132 |
+
|
| 133 |
+
if "raw_output" in criteria:
|
| 134 |
+
print(criteria["raw_output"])
|
| 135 |
+
else:
|
| 136 |
+
if "measure" in criteria:
|
| 137 |
+
print(f"\n📋 Measure: {criteria['measure']}")
|
| 138 |
+
|
| 139 |
+
if "required_tests" in criteria:
|
| 140 |
+
print(f"\n🔬 Required Tests/Procedures:")
|
| 141 |
+
if isinstance(criteria['required_tests'], list):
|
| 142 |
+
for test in criteria['required_tests']:
|
| 143 |
+
print(f" • {test}")
|
| 144 |
+
else:
|
| 145 |
+
print(f" {criteria['required_tests']}")
|
| 146 |
+
|
| 147 |
+
if "required_medications" in criteria:
|
| 148 |
+
print(f"\n💊 Required Medications:")
|
| 149 |
+
if isinstance(criteria['required_medications'], list):
|
| 150 |
+
for med in criteria['required_medications']:
|
| 151 |
+
print(f" • {med}")
|
| 152 |
+
else:
|
| 153 |
+
print(f" {criteria['required_medications']}")
|
| 154 |
+
|
| 155 |
+
if "codes" in criteria:
|
| 156 |
+
print(f"\n🏷️ Relevant Codes:")
|
| 157 |
+
if isinstance(criteria['codes'], dict):
|
| 158 |
+
for code_type, codes in criteria['codes'].items():
|
| 159 |
+
print(f" {code_type}: {codes}")
|
| 160 |
+
else:
|
| 161 |
+
print(f" {criteria['codes']}")
|
| 162 |
+
|
| 163 |
+
if "timeframes" in criteria:
|
| 164 |
+
print(f"\n⏰ Timeframes:")
|
| 165 |
+
if isinstance(criteria['timeframes'], dict):
|
| 166 |
+
for requirement, timeframe in criteria['timeframes'].items():
|
| 167 |
+
print(f" {requirement}: {timeframe}")
|
| 168 |
+
else:
|
| 169 |
+
print(f" {criteria['timeframes']}")
|
| 170 |
+
|
| 171 |
+
if "inclusions" in criteria:
|
| 172 |
+
print(f"\n✅ Inclusion Criteria:")
|
| 173 |
+
if isinstance(criteria['inclusions'], list):
|
| 174 |
+
for inclusion in criteria['inclusions']:
|
| 175 |
+
print(f" • {inclusion}")
|
| 176 |
+
else:
|
| 177 |
+
print(f" {criteria['inclusions']}")
|
| 178 |
+
|
| 179 |
+
if "exclusions" in criteria:
|
| 180 |
+
print(f"\n❌ Exclusion Criteria:")
|
| 181 |
+
if isinstance(criteria['exclusions'], list):
|
| 182 |
+
for exclusion in criteria['exclusions']:
|
| 183 |
+
print(f" • {exclusion}")
|
| 184 |
+
else:
|
| 185 |
+
print(f" {criteria['exclusions']}")
|
| 186 |
+
|
| 187 |
+
def get_user_feedback(self) -> tuple[str, str]:
|
| 188 |
+
"""
|
| 189 |
+
Get user feedback on the displayed criteria.
|
| 190 |
+
Returns (action, suggestions) where action is 'continue', 'modify', or 'exit'
|
| 191 |
+
"""
|
| 192 |
+
print(f"\n{'='*60}")
|
| 193 |
+
print("REVIEW THE CRITERIA ABOVE")
|
| 194 |
+
print(f"{'='*60}")
|
| 195 |
+
print("Options:")
|
| 196 |
+
print("• Press ENTER to continue with these criteria")
|
| 197 |
+
print("• Type suggestions to modify the criteria")
|
| 198 |
+
print("• Type 'exit' to stop the analysis")
|
| 199 |
+
print(f"{'='*60}")
|
| 200 |
+
|
| 201 |
+
user_input = input("\nYour choice: ").strip()
|
| 202 |
+
|
| 203 |
+
if user_input.lower() == 'exit':
|
| 204 |
+
return 'exit', ''
|
| 205 |
+
elif user_input == '':
|
| 206 |
+
return 'continue', ''
|
| 207 |
+
else:
|
| 208 |
+
return 'modify', user_input
|
| 209 |
+
|
| 210 |
+
def run_interactive(self) -> str:
|
| 211 |
+
"""
|
| 212 |
+
Run the interactive HEDIS compliance analysis with user feedback loop.
|
| 213 |
+
"""
|
| 214 |
+
print(f"Starting Interactive HEDIS Analysis for measure {self.measure_code}")
|
| 215 |
+
|
| 216 |
+
suggestions = ""
|
| 217 |
+
while True:
|
| 218 |
+
# Get criteria from expert
|
| 219 |
+
print(f"\n🔍 Getting HEDIS criteria for {self.measure_code}...")
|
| 220 |
+
criteria = self.get_hedis_criteria_with_feedback(suggestions)
|
| 221 |
+
|
| 222 |
+
# Display criteria to user
|
| 223 |
+
self.display_criteria(criteria)
|
| 224 |
+
|
| 225 |
+
# Get user feedback
|
| 226 |
+
action, user_suggestions = self.get_user_feedback()
|
| 227 |
+
|
| 228 |
+
if action == 'exit':
|
| 229 |
+
print("\n❌ Analysis cancelled by user.")
|
| 230 |
+
return "Analysis cancelled by user."
|
| 231 |
+
elif action == 'continue':
|
| 232 |
+
print("\n✅ Criteria approved! Proceeding with chart analysis...")
|
| 233 |
+
self.hedis_criteria = criteria
|
| 234 |
+
break
|
| 235 |
+
elif action == 'modify':
|
| 236 |
+
print(f"\n🔄 Modifying criteria based on your suggestions...")
|
| 237 |
+
suggestions = user_suggestions
|
| 238 |
+
continue
|
| 239 |
+
|
| 240 |
+
# Now proceed with the rest of the analysis using approved criteria
|
| 241 |
+
return self.run_full_analysis()
|
| 242 |
+
|
| 243 |
+
def run_full_analysis(self) -> str:
|
| 244 |
+
"""
|
| 245 |
+
Run the full analysis with approved criteria.
|
| 246 |
+
"""
|
| 247 |
+
# Create tasks for patient analysis and reporting
|
| 248 |
+
patient_task = Task(
|
| 249 |
+
description=(
|
| 250 |
+
f"Using the patient_chart_search tool, extract only the tests, diagnoses, procedures, "
|
| 251 |
+
f"medications, and dates relevant to {self.measure_code}. Only include measure related info from the patient charts. "
|
| 252 |
+
f"Use these approved criteria as reference: {json.dumps(self.hedis_criteria, indent=2)}"
|
| 253 |
+
),
|
| 254 |
+
expected_output="Structured list of patient evidence with dates.",
|
| 255 |
+
agent=self.patient_analyzer,
|
| 256 |
+
async_execution=True,
|
| 257 |
+
)
|
| 258 |
+
|
| 259 |
+
chart_analysis_task = Task(
|
| 260 |
description=(
|
| 261 |
+
f"Create a comprehensive markdown report analyzing the patient chart for HEDIS {self.measure_code} "
|
| 262 |
+
f"for measurement year {self.measurement_year}.\n\n"
|
| 263 |
+
"Sections:\n"
|
| 264 |
+
"### HEDIS Measure Criteria (Approved)\n"
|
| 265 |
+
"### Diagnoses Found\n"
|
| 266 |
+
"### Procedures Found\n"
|
| 267 |
+
"### Laboratory Tests\n"
|
| 268 |
+
"### Medications\n"
|
| 269 |
+
"### Vital Signs\n"
|
| 270 |
+
"### Exclusions Found\n"
|
| 271 |
+
"### Missing Information\n"
|
| 272 |
+
"### Compliance Summary\n"
|
| 273 |
+
f"Use these approved criteria as reference: {json.dumps(self.hedis_criteria, indent=2)}\n"
|
| 274 |
+
"Make sure only measure related information is included."
|
| 275 |
+
),
|
| 276 |
+
expected_output="A thorough markdown report of HEDIS findings with compliance assessment.",
|
| 277 |
+
agent=self.hedis_chart_reporter,
|
| 278 |
+
context=[patient_task],
|
| 279 |
+
)
|
| 280 |
+
|
| 281 |
+
# Create crew for full analysis
|
| 282 |
+
analysis_crew = Crew(
|
| 283 |
+
agents=[self.patient_analyzer, self.hedis_chart_reporter],
|
| 284 |
+
tasks=[patient_task, chart_analysis_task],
|
| 285 |
+
process=Process.sequential,
|
| 286 |
+
verbose=True,
|
| 287 |
+
)
|
| 288 |
+
|
| 289 |
+
# Execute full analysis
|
| 290 |
+
inputs = {
|
| 291 |
+
"measure_code": self.measure_code,
|
| 292 |
+
"measurement_year": self.measurement_year,
|
| 293 |
+
}
|
| 294 |
+
|
| 295 |
+
print(f"\n🔄 Running patient chart analysis...")
|
| 296 |
+
result = analysis_crew.kickoff(inputs=inputs)
|
| 297 |
+
|
| 298 |
+
print(f"\n✅ Analysis complete!")
|
| 299 |
+
return str(result)
|
| 300 |
+
|
| 301 |
+
def run_non_interactive(self) -> str:
|
| 302 |
+
"""
|
| 303 |
+
Run the original non-interactive analysis (for backwards compatibility).
|
| 304 |
+
"""
|
| 305 |
+
hedis_task = Task(
|
| 306 |
+
description=(
|
| 307 |
+
f"Provide the official criteria for HEDIS measure {self.measure_code}, including:\n"
|
| 308 |
"- Required tests or procedures\n"
|
| 309 |
"- Diagnosis/procedure codes\n"
|
| 310 |
"- Required medications\n"
|
| 311 |
"- Inclusion and exclusion criteria\n"
|
| 312 |
+
f"- Timeframes for each requirement (e.g., within X years of {self.measurement_year})\n\n"
|
| 313 |
+
f"You MUST compute and include absolute date ranges using {self.measurement_year}.\n"
|
| 314 |
"Return JSON with fields: 'measure','required_tests','required_medications',"
|
| 315 |
"'codes','timeframes','inclusions','exclusions'."
|
| 316 |
),
|
|
|
|
| 319 |
async_execution=True,
|
| 320 |
)
|
| 321 |
|
| 322 |
+
patient_task = Task(
|
| 323 |
description=(
|
| 324 |
+
f"Using the patient_chart_search tool, extract only the tests, diagnoses, procedures, "
|
| 325 |
+
f"medications, and dates relevant to {self.measure_code}. Only include measure related info from the patient charts"
|
| 326 |
),
|
| 327 |
expected_output="Structured list of patient evidence with dates.",
|
| 328 |
agent=self.patient_analyzer,
|
| 329 |
async_execution=True,
|
| 330 |
)
|
| 331 |
|
| 332 |
+
chart_analysis_task = Task(
|
| 333 |
description=(
|
| 334 |
+
f"Create a comprehensive markdown report analyzing the patient chart for HEDIS {self.measure_code} "
|
| 335 |
+
f"for measurement year {self.measurement_year}.\n\n"
|
| 336 |
"Sections:\n"
|
| 337 |
"### Diagnoses Found\n"
|
| 338 |
"### Procedures Found\n"
|
|
|
|
| 346 |
),
|
| 347 |
expected_output="A thorough markdown report of HEDIS findings.",
|
| 348 |
agent=self.hedis_chart_reporter,
|
| 349 |
+
context=[hedis_task, patient_task],
|
| 350 |
)
|
| 351 |
|
| 352 |
+
crew = Crew(
|
| 353 |
agents=[self.hedis_expert, self.patient_analyzer, self.hedis_chart_reporter],
|
| 354 |
+
tasks=[hedis_task, patient_task, chart_analysis_task],
|
| 355 |
process=Process.sequential,
|
| 356 |
verbose=True,
|
| 357 |
)
|
| 358 |
|
|
|
|
| 359 |
inputs = {
|
| 360 |
"measure_code": self.measure_code,
|
| 361 |
"measurement_year": self.measurement_year,
|
| 362 |
}
|
| 363 |
+
result = crew.kickoff(inputs=inputs)
|
| 364 |
+
return str(result)
|
| 365 |
+
|
| 366 |
+
def run(self) -> str:
|
| 367 |
+
"""
|
| 368 |
+
Entry point - runs non-interactive analysis by default for app compatibility.
|
| 369 |
+
Use run_interactive() for interactive mode.
|
| 370 |
+
"""
|
| 371 |
+
return self.run_non_interactive()
|
| 372 |
+
|
| 373 |
+
|
| 374 |
+
# Usage example:
|
| 375 |
+
"""
|
| 376 |
+
# Initialize the engine
|
| 377 |
+
vectordb = your_chroma_db # Your Chroma vector database
|
| 378 |
+
engine = InteractiveHedisComplianceEngine(
|
| 379 |
+
vectordb=vectordb,
|
| 380 |
+
measure_code="CDC-A1C",
|
| 381 |
+
measurement_year=2024
|
| 382 |
+
)
|
| 383 |
+
|
| 384 |
+
# Run interactive analysis
|
| 385 |
+
result = engine.run()
|
| 386 |
+
print(result)
|
| 387 |
+
"""
|