pyregularexpression 0.1.0

Collection of reusable regular‑expression patterns

pyRegularExpression

This package provides a collection of regular expression-based functions to identify and extract components of the scientific process from text. These components include things like identifying if a text discusses adherence, compliance, eligibility criteria, and more.

Installation

pip install pyregularexpression

Available Finder Modules

This package contains a number of finder modules, each designed to find a specific concept in a text. Each module contains one or more functions that implement different versions of a regular expression with varying levels of precision and recall.

Below is a list of the available finder modules and their purpose:

• Adherence Compliance (adherence_compliance_finder.py): adherence_compliance_finder.py – precision/recall ladder for treatment adherence / compliance metrics.
  • find_adherence_compliance_v1(text)
  • find_adherence_compliance_v2(text, window=4)
  • find_adherence_compliance_v3(text, block_chars=400)
  • find_adherence_compliance_v4(text, window=12)
  • find_adherence_compliance_v5(text)
• Algorithm Validation (algorithm_validation_finder.py): algorithm_validation_finder.py – precision/recall ladder for algorithm validation statements.
  • find_algorithm_validation_v1(text)
  • find_algorithm_validation_v2(text, window=4)
  • find_algorithm_validation_v3(text, block_chars=400)
  • find_algorithm_validation_v4(text, window=12)
  • find_algorithm_validation_v5(text)
• Allocation Concealment (allocation_concealment_finder.py): allocation_concealment_finder.py – precision/recall ladder for allocation concealment methods.
  • find_allocation_concealment_v1(text)
  • find_allocation_concealment_v2(text, window=4)
  • find_allocation_concealment_v3(text, block_chars=400)
  • find_allocation_concealment_v4(text, window=12)
  • find_allocation_concealment_v5(text)
• Attrition Criteria (attrition_criteria_finder.py): attrition_criteria_finder.py – precision/recall ladder for attrition criteria (post‑enrolment loss).
  • find_attrition_criteria_v1(text)
  • find_attrition_criteria_v2(text, window=4)
  • find_attrition_criteria_v3(text, block_chars=400)
  • find_attrition_criteria_v4(text, window=12)
  • find_attrition_criteria_v5(text)
• Background Rationale (background_rationale_finder.py): background_rationale_finder.py – precision/recall ladder for study background / rationale statements.
  • find_background_rationale_v1(text)
  • find_background_rationale_v2(text, window=4)
  • find_background_rationale_v3(text, block_chars=500)
  • find_background_rationale_v4(text, window=12)
  • find_background_rationale_v5(text)
• Baseline Data (baseline_data_finder.py): baseline_data_finder.py – precision/recall ladder for baseline participant characteristics.
  • find_baseline_data_v1(text)
  • find_baseline_data_v2(text, window=4)
  • find_baseline_data_v3(text, block_chars=400)
  • find_baseline_data_v4(text, window=12)
  • find_baseline_data_v5(text)
• Blinding Masking (blinding_masking_finder.py): blinding_masking_finder.py – precision/recall ladder for blinding / masking status.
  • find_blinding_masking_v1(text)
  • find_blinding_masking_v2(text, window=4)
  • find_blinding_masking_v3(text, block_chars=400)
  • find_blinding_masking_v4(text, window=12)
  • find_blinding_masking_v5(text)
• Changes To Outcomes (changes_to_outcomes_finder.py): changes_to_outcomes_finder.py – precision/recall ladder for changes to prespecified outcomes after trial initiation.
  • find_changes_to_outcomes_v1(text)
  • find_changes_to_outcomes_v2(text, window=4)
  • find_changes_to_outcomes_v3(text, block_chars=400)
  • find_changes_to_outcomes_v4(text, window=12)
  • find_changes_to_outcomes_v5(text)
• Comparator Cohort (comparator_cohort_finder.py): comparator_cohort_finder.py – precision/recall ladder for comparator (control) cohort statements.
  • find_comparator_cohort_v1(text)
  • find_comparator_cohort_v2(text, window=4)
  • find_comparator_cohort_v3(text, block_chars=400)
  • find_comparator_cohort_v4(text, window=12)
  • find_comparator_cohort_v5(text)
• Competing Risk Analysis (competing_risk_analysis_finder.py): competing_risk_analysis_finder.py – precision/recall ladder for competing‑risk analyses.
  • find_competing_risk_analysis_v1(text)
  • find_competing_risk_analysis_v2(text, window=4)
  • find_competing_risk_analysis_v3(text, block_chars=400)
  • find_competing_risk_analysis_v4(text, window=12)
  • find_competing_risk_analysis_v5(text)
• Conflict Of Interest (conflict_of_interest_finder.py): conflict_of_interest_finder.py – precision/recall ladder for conflict‑of‑interest disclosures.
  • find_conflict_of_interest_v1(text)
  • find_conflict_of_interest_v2(text, window=4)
  • find_conflict_of_interest_v3(text, block_chars=400)
  • find_conflict_of_interest_v4(text, window=12)
  • find_conflict_of_interest_v5(text)
• Covariate Adjustment (covariate_adjustment_finder.py): covariate_adjustment_finder.py – precision/recall ladder for covariate adjustment statements.
  • find_covariate_adjustment_v1(text)
  • find_covariate_adjustment_v2(text, window=4)
  • find_covariate_adjustment_v3(text, block_chars=300)
  • find_covariate_adjustment_v4(text, window=6)
  • find_covariate_adjustment_v5(text)
• Data Access (data_access_finder.py): data_access_finder.py – precision/recall ladder for data‑access / availability statements.
  • find_data_access_v1(text)
  • find_data_access_v2(text, window=3)
  • find_data_access_v3(text, block_chars=400)
  • find_data_access_v4(text, window=12)
  • find_data_access_v5(text)
• Data Linkage Method (data_linkage_method_finder.py): data_linkage_method_finder.py – precision/recall ladder for data‑linkage methods.
  • find_data_linkage_method_v1(text)
  • find_data_linkage_method_v2(text, window=3)
  • find_data_linkage_method_v3(text, block_chars=400)
  • find_data_linkage_method_v4(text, window=12)
  • find_data_linkage_method_v5(text)
• Data Provenance (data_provenance_finder.py): data_provenance_finder.py – precision/recall ladder for data provenance
  • find_data_provenance_v1(text)
  • find_data_provenance_v2(text, window=4)
  • find_data_provenance_v3(text, block_chars=400)
  • find_data_provenance_v4(text, window=12)
  • find_data_provenance_v5(text)
• Data Safety Monitoring (data_safety_monitoring_finder.py): data_safety_monitoring_finder.py – precision/recall ladder for Data‑Safety Monitoring descriptions.
  • find_data_safety_monitoring_v1(text)
  • find_data_safety_monitoring_v2(text, window=4)
  • find_data_safety_monitoring_v3(text, block_chars=400)
  • find_data_safety_monitoring_v4(text, window=12)
  • find_data_safety_monitoring_v5(text)
• Data Sharing Statement (data_sharing_statement_finder.py): data_sharing_statement_finder.py – precision/recall ladder for data‑sharing statements.
  • find_data_sharing_statement_v1(text)
  • find_data_sharing_statement_v2(text, window=4)
  • find_data_sharing_statement_v3(text, block_chars=400)
  • find_data_sharing_statement_v4(text, window=12)
  • find_data_sharing_statement_v5(text)
• Data Source Type (data_source_type_finder.py): data_source_type_finder.py – precision/recall ladder for data‑source type declarations.
  • find_data_source_type_v1(text)
  • find_data_source_type_v2(text, window=2)
  • find_data_source_type_v3(text, block_chars=400)
  • find_data_source_type_v4(text, window=12)
  • find_data_source_type_v5(text)
• Demographic Restriction (demographic_restriction_finder.py): demographic_restriction_finder.py – precision/recall ladder for demographic‑restriction statements.
  • find_demographic_restriction_v1(text)
  • find_demographic_restriction_v2(text, window=4)
  • find_demographic_restriction_v3(text, block_chars=400)
  • find_demographic_restriction_v4(text, window=12)
  • find_demographic_restriction_v5(text)
• Dose Response Analysis (dose_response_analysis_finder.py): dose_response_analysis_finder.py – precision/recall ladder for dose‑response / exposure‑response analyses.
  • find_dose_response_analysis_v1(text)
  • find_dose_response_analysis_v2(text, window=4)
  • find_dose_response_analysis_v3(text, block_chars=400)
  • find_dose_response_analysis_v4(text, window=12)
  • find_dose_response_analysis_v5(text)
• Eligibility Criteria (eligibility_criteria_finder.py): eligibility_criteria_finder.py – precision/recall ladder for inclusion / exclusion eligibility criteria statements.
  • find_eligibility_criteria_v1(text)
  • find_eligibility_criteria_v2(text, window=4)
  • find_eligibility_criteria_v3(text, block_chars=500)
  • find_eligibility_criteria_v4(text, window=12)
  • find_eligibility_criteria_v5(text)
• Entry Event (entry_event_finder.py): entry_event_finder.py – precision/recall ladder for entry‑event statements.
  • find_entry_event_v1(text)
  • find_entry_event_v2(text, window=4)
  • find_entry_event_v3(text, block_chars=400)
  • find_entry_event_v4(text, window=12)
  • find_entry_event_v5(text)
• Ethics Approval (ethics_approval_finder.py): ethics_approval_finder.py – precision/recall ladder for ethics approval & consent statements.
  • find_ethics_approval_v1(text)
  • find_ethics_approval_v2(text, window=4)
  • find_ethics_approval_v3(text, block_chars=400)
  • find_ethics_approval_v4(text, window=12)
  • find_ethics_approval_v5(text)
• Event Adjudication (event_adjudication_finder.py): event_adjudication_finder.py – precision/recall ladder for event‑adjudication descriptions.
  • find_event_adjudication_v1(text)
  • find_event_adjudication_v2(text, window=5)
  • find_event_adjudication_v3(text, block_chars=400)
  • find_event_adjudication_v4(text, window=12)
  • find_event_adjudication_v5(text)
• Exclusion Rule (exclusion_rule_finder.py): exclusion_rule_finder.py – precision/recall ladder for exclusion‑rule statements.
  • find_exclusion_rule_v1(text)
  • find_exclusion_rule_v2(text, window=4)
  • find_exclusion_rule_v3(text, block_chars=400)
  • find_exclusion_rule_v4(text, window=12)
  • find_exclusion_rule_v5(text)
• Exit Criterion (exit_criterion_finder.py): exit_criterion_finder.py – precision/recall ladder for exit-criterion statements.
  • find_exit_criterion_v1(text)
  • find_exit_criterion_v2(text, window=4)
  • find_exit_criterion_v3(text, block_chars=400)
  • find_exit_criterion_v4(text, window=12)
  • find_exit_criterion_v5(text)
• Exposure Definition (exposure_definition_finder.py): exposure_definition_finder.py – precision/recall ladder for exposure definition statements.
  • find_exposure_definition_v1(text)
  • find_exposure_definition_v2(text, window=4)
  • find_exposure_definition_v3(text, block_chars=400)
  • find_exposure_definition_v4(text, window=12)
  • find_exposure_definition_v5(text)
• Follow Up Period (follow_up_period_finder.py): follow_up_period_finder.py – precision/recall ladder for follow‑up period definitions.
  • find_follow_up_period_v1(text)
  • find_follow_up_period_v2(text, window=4)
  • find_follow_up_period_v3(text, block_chars=400)
  • find_follow_up_period_v4(text, window=12)
  • find_follow_up_period_v5(text)
• Funding Statement (funding_statement_finder.py): funding_statement_finder.py – precision/recall ladder for study funding statements.
  • find_funding_statement_v1(text)
  • find_funding_statement_v2(text, window=4)
  • find_funding_statement_v3(text, block_chars=400)
  • find_funding_statement_v4(text, window=12)
  • find_funding_statement_v5(text)
• Generalizability (generalizability_finder.py): generalizability_finder.py – precision/recall ladder for generalizability / external validity statements.
  • find_generalizability_v1(text)
  • find_generalizability_v2(text, window=4)
  • find_generalizability_v3(text, block_chars=400)
  • find_generalizability_v4(text, window=12)
  • find_generalizability_v5(text)
• Harms Adverse Event (harms_adverse_event_finder.py): harms_adverse_event_finder.py – precision/recall ladder for harms / adverse events.
  • find_harms_adverse_event_v1(text)
  • find_harms_adverse_event_v2(text, window=4)
  • find_harms_adverse_event_v3(text, block_chars=400)
  • find_harms_adverse_event_v4(text, window=12)
  • find_harms_adverse_event_v5(text)
• Healthcare Setting (healthcare_setting_finder.py): healthcare_setting_finder.py – precision/recall ladder for health‑care setting statements.
  • find_healthcare_setting_v1(text)
  • find_healthcare_setting_v2(text, window=3)
  • find_healthcare_setting_v3(text, block_chars=400)
  • find_healthcare_setting_v4(text, window=12)
  • find_healthcare_setting_v5(text)
• Inclusion Rule (inclusion_rule_finder.py): inclusion_rule_finder.py – precision/recall ladder for inclusion‑rule statements.
  • find_inclusion_rule_v1(text)
  • find_inclusion_rule_v2(text, window=4)
  • find_inclusion_rule_v3(text, block_chars=400)
  • find_inclusion_rule_v4(text, window=12)
  • find_inclusion_rule_v5(text)
• Index Date (index_date_finder.py): index_date_finder.py – precision/recall ladder for index‑date definition statements.
  • find_index_date_v1(text)
  • find_index_date_v2(text, window=4)
  • find_index_date_v3(text, block_chars=400)
  • find_index_date_v4(text, window=12)
  • find_index_date_v5(text)
• Interim Analysis Stopping Rules (interim_analysis_stopping_rules_finder.py): interim_analysis_stopping_rules_finder.py – multi-tiered finder for interim analysis stopping rules.
  • find_stopping_rule_v1(text)
  • find_stopping_rule_v2(text)
  • find_stopping_rule_v3(text)
• Interventions (interventions_finder.py): interventions_finder.py – precision/recall ladder for interventions / treatments delivered to study arms.
  • find_interventions_v1(text)
  • find_interventions_v2(text, window=4)
  • find_interventions_v3(text, block_chars=400)
  • find_interventions_v4(text, window=12)
  • find_interventions_v5(text)
• Limitations (limitations_finder.py): limitations_finder.py – precision/recall ladder for study limitations sections.
  • find_limitations_v1(text)
  • find_limitations_v2(text, window=4)
  • find_limitations_v3(text, block_chars=400)
  • find_limitations_v4(text, window=12)
  • find_limitations_v5(text)
• Losses Exclusion (losses_exclusion_finder.py): losses_exclusion_finder.py – precision/recall ladder for losses and exclusions after allocation.
  • find_losses_exclusion_v1(text)
  • find_losses_exclusion_v2(text, window=4)
  • find_losses_exclusion_v3(text, block_chars=500)
  • find_losses_exclusion_v4(text, window=12)
  • find_losses_exclusion_v5(text)
• Medical Code (medical_code_finder.py): medical_code_finder.py – precision/recall ladder for medical code statements.
  • find_medical_code_v1(text)
  • find_medical_code_v2(text, window=5)
  • find_medical_code_v3(text, block_chars=300)
  • find_medical_code_v4(text)
  • find_medical_code_v5(text)
• Missing Data Handling (missing_data_handling_finder.py): missing_data_handling_finder.py – precision/recall ladder for missing‑data handling methods.
  • find_missing_data_handling_v1(text)
  • find_missing_data_handling_v2(text, window=4)
  • find_missing_data_handling_v3(text, block_chars=400)
  • find_missing_data_handling_v4(text, window=12)
  • find_missing_data_handling_v5(text)
• Numbers Analyzed (numbers_analyzed_finder.py): numbers_analyzed_finder.py – precision/recall ladder for numbers analysed in each analysis population.
  • find_numbers_analyzed_v1(text)
  • find_numbers_analyzed_v2(text, window=4)
  • find_numbers_analyzed_v3(text, block_chars=400)
  • find_numbers_analyzed_v4(text, window=12)
  • find_numbers_analyzed_v5(text)
• Objective Hypothesis (objective_hypothesis_finder.py): objective_hypothesis_finder.py – precision/recall ladder for study objectives / hypotheses statements.
  • find_objective_hypothesis_v1(text)
  • find_objective_hypothesis_v2(text, window=3)
  • find_objective_hypothesis_v3(text, block_chars=400)
  • find_objective_hypothesis_v4(text, window=12)
  • find_objective_hypothesis_v5(text)
• Outcome Ascertainment (outcome_ascertainment_finder.py): outcome_ascertainment_finder.py – precision/recall ladder for outcome ascertainment statements.
  • find_outcome_ascertainment_v1(text)
  • find_outcome_ascertainment_v2(text, window=4)
  • find_outcome_ascertainment_v3(text, block_chars=400)
  • find_outcome_ascertainment_v4(text, window=12)
  • find_outcome_ascertainment_v5(text)
• Outcome Definition (outcome_definition_finder.py): outcome_definition_finder.py – precision/recall ladder for outcome definition statements.
  • find_outcome_definition_v1(text)
  • find_outcome_definition_v2(text, window=4)
  • find_outcome_definition_v3(text, block_chars=400)
  • find_outcome_definition_v4(text, window=12)
  • find_outcome_definition_v5(text)
• Outcome Endpoints (outcome_endpoints_finder.py): outcome_endpoints_finder.py – precision/recall ladder for primary and secondary outcomes / endpoints statements.
  • find_outcome_endpoints_v1(text)
  • find_outcome_endpoints_v2(text, window=4)
  • find_outcome_endpoints_v3(text, block_chars=500)
  • find_outcome_endpoints_v4(text, window=12)
  • find_outcome_endpoints_v5(text)
• Participant Flow (participant_flow_finder.py): participant_flow_finder.py – precision/recall ladder for participant flow statements (CONSORT flow).
  • find_participant_flow_v1(text)
  • find_participant_flow_v2(text, window=4)
  • find_participant_flow_v3(text, block_chars=600)
  • find_participant_flow_v4(text, window=12)
  • find_participant_flow_v5(text)
• Propensity Score Method (propensity_score_method_finder.py): propensity_score_method_finder.py – precision/recall ladder for propensity-score methods.
  • find_propensity_score_method_v1(text)
  • find_propensity_score_method_v2(text, window=4)
  • find_propensity_score_method_v3(text, block_chars=400)
  • find_propensity_score_method_v4(text, window=12)
  • find_propensity_score_method_v5(text)
• Random Sequence Generation (random_sequence_generation_finder.py): random_sequence_generation_finder.py – precision/recall ladder for random allocation-sequence generation methods.
  • find_random_sequence_generation_v1(text)
  • find_random_sequence_generation_v2(text, window=4)
  • find_random_sequence_generation_v3(text, block_chars=400)
  • find_random_sequence_generation_v4(text, window=12)
  • find_random_sequence_generation_v5(text)
• Randomization Implementation (randomization_implementation_finder.py): randomization_implementation_finder.py – precision/recall ladder for randomization implementation (who generated sequence, who enrolled, who assigned).
  • find_randomization_implementation_v1(text)
  • find_randomization_implementation_v2(text, window=4)
  • find_randomization_implementation_v3(text, block_chars=500)
  • find_randomization_implementation_v4(text, window=12)
  • find_randomization_implementation_v5(text)
• Randomization Type Restriction (randomization_type_restriction_finder.py): randomization_type_restriction_finder.py – precision/recall ladder for randomization type / restrictions (blocking, stratification, ratio).
  • find_randomization_type_restriction_v1(text)
  • find_randomization_type_restriction_v2(text, window=4)
  • find_randomization_type_restriction_v3(text, block_chars=400)
  • find_randomization_type_restriction_v4(text, window=12)
  • find_randomization_type_restriction_v5(text)
• Recruitment Timeline (recruitment_timeline_finder.py): recruitment_timeline_finder.py – precision/recall ladder for recruitment period / timeline.
  • find_recruitment_timeline_v1(text)
  • find_recruitment_timeline_v2(text, window=4)
  • find_recruitment_timeline_v3(text, block_chars=500)
  • find_recruitment_timeline_v4(text, window=12)
  • find_recruitment_timeline_v5(text)
• Risk Of Bias Assessment (risk_of_bias_assessment_finder.py): risk_of_bias_assessment_finder.py – precision/recall ladder for risk‑of‑bias assessments in systematic reviews.
  • find_risk_of_bias_assessment_v1(text)
  • find_risk_of_bias_assessment_v2(text, window=4)
  • find_risk_of_bias_assessment_v3(text, block_chars=400)
  • find_risk_of_bias_assessment_v4(text, window=12)
  • find_risk_of_bias_assessment_v5(text)
• Sensitivity Analysis (sensitivity_analysis_finder.py): sensitivity_analysis_finder.py – precision/recall ladder for sensitivity analysis statements.
  • find_sensitivity_analysis_v1(text)
  • find_sensitivity_analysis_v2(text, window=4)
  • find_sensitivity_analysis_v3(text, block_chars=400)
  • find_sensitivity_analysis_v4(text, window=12)
  • find_sensitivity_analysis_v5(text)
• Settings Locations (settings_locations_finder.py): settings_locations_finder.py – precision/recall ladder for study settings / locations
  • find_settings_locations_v1(text)
  • find_settings_locations_v2(text, window=4)
  • find_settings_locations_v3(text, block_chars=400)
  • find_settings_locations_v4(text, window=12)
  • find_settings_locations_v5(text)
• Severity Definition (severity_definition_finder.py): severity_definition_finder.py – precision/recall ladder for severity definition statements.
  • find_severity_definition_v1(text)
  • find_severity_definition_v2(text, window=4)
  • find_severity_definition_v3(text, block_chars=400)
  • find_severity_definition_v4(text, window=12)
  • find_severity_definition_v5(text)
• Similarity Of Interventions (similarity_of_interventions_finder.py): similarity_of_interventions_finder.py – precision/recall ladder for similarity of interventions.
  • find_similarity_of_interventions_v1(text)
  • find_similarity_of_interventions_v2(text, window=4)
  • find_similarity_of_interventions_v3(text, block_chars=400)
  • find_similarity_of_interventions_v4(text, window=12)
  • find_similarity_of_interventions_v5(text)
• Statistical Analysis Additional Method (statistical_analysis_additional_method_finder.py): statistical_analysis_additional_method_finder.py – precision/recall ladder for statistical methods of additional analyses (secondary, subgroup, exploratory).
  • find_statistical_analysis_additional_method_v1(text)
  • find_statistical_analysis_additional_method_v2(text, window=4)
  • find_statistical_analysis_additional_method_v3(text, block_chars=500)
  • find_statistical_analysis_additional_method_v4(text, window=12)
  • find_statistical_analysis_additional_method_v5(text)
• Statistical Analysis (statistical_analysis_finder.py): statistical_analysis_finder.py – precision/recall ladder for statistical analysis statements.
  • find_statistical_analysis_v1(text)
  • find_statistical_analysis_v2(text, window=4)
  • find_statistical_analysis_v3(text, block_chars=300)
  • find_statistical_analysis_v4(text, window=6)
  • find_statistical_analysis_v5(text)
• Statistical Analysis Primary Analysis (statistical_analysis_primary_analysis_finder.py): statistical_analysis_primary_analysis_finder.py – precision/recall ladder for statistical methods of the primary analysis.
  • find_statistical_analysis_primary_analysis_v1(text)
  • find_statistical_analysis_primary_analysis_v2(text, window=4)
  • find_statistical_analysis_primary_analysis_v3(text, block_chars=500)
  • find_statistical_analysis_primary_analysis_v4(text, window=12)
  • find_statistical_analysis_primary_analysis_v5(text)
• Study Design (study_design_finder.py): study_design_finder.py – precision/recall ladder for study design declarations.
  • find_study_design_v1(text)
  • find_study_design_v2(text, window=4)
  • find_study_design_v3(text, block_chars=400)
  • find_study_design_v4(text, window=12)
  • find_study_design_v5(text)
• Study Period (study_period_finder.py): study_period_finder.py – precision/recall ladder for study period calendar windows.
  • find_study_period_v1(text)
  • find_study_period_v2(text, window=4)
  • find_study_period_v3(text, block_chars=400)
  • find_study_period_v4(text, window=12)
  • find_study_period_v5(text)
• Subgroup Analysis (subgroup_analysis_finder.py): subgroup_analysis_finder.py – precision/recall ladder for subgroup / interaction analyses.
  • find_subgroup_analysis_v1(text)
  • find_subgroup_analysis_v2(text, window=4)
  • find_subgroup_analysis_v3(text, block_chars=400)
  • find_subgroup_analysis_v4(text, window=12)
  • find_subgroup_analysis_v5(text)
• Treatment Definition (treatment_definition_finder.py): treatment_definition_finder.py – precision/recall ladder for treatment definition statements.
  • find_treatment_definition_v1(text)
  • find_treatment_definition_v2(text, window=4)
  • find_treatment_definition_v3(text, block_chars=400)
  • find_treatment_definition_v4(text, window=12)
  • find_treatment_definition_v5(text)
• Trial Design Changes (trial_design_changes_finder.py): trial_design_changes_finder.py – precision/recall ladder for changes to trial design/protocol after initiation.
  • find_trial_design_changes_v1(text)
  • find_trial_design_changes_v2(text, window=4)
  • find_trial_design_changes_v3(text, block_chars=400)
  • find_trial_design_changes_v4(text, window=12)
  • find_trial_design_changes_v5(text)
• Trial Design (trial_design_finder.py): trial_design_finder.py – precision/recall ladder for clinical/epidemiological trial or study design statements.
  • find_trial_design_v1(text)
  • find_trial_design_v2(text, window=4)
  • find_trial_design_v3(text, block_chars=400)
  • find_trial_design_v4(text, window=12)
  • find_trial_design_v5(text)
• Trial Registration (trial_registration_finder.py): trial_registration_finder.py – precision/recall ladder for prospective trial registration statements.
  • find_trial_registration_v1(text)
  • find_trial_registration_v2(text, window=4)
  • find_trial_registration_v3(text, block_chars=400)
  • find_trial_registration_v4(text, window=12)
  • find_trial_registration_v5(text)
• Washout Period (washout_period_finder.py): washout_period_finder.py – precision/recall ladder for washout period definitions.
  • find_washout_period_v1(text)
  • find_washout_period_v2(text, window=4)
  • find_washout_period_v3(text, block_chars=400)
  • find_washout_period_v4(text, window=12)
  • find_washout_period_v5(text)

Usage

Finder Functions

Each *_finder.py module contains finder functions that return a list of tuples, where each tuple contains the start and end character indices of a match and the matched string itself.

For example, the adherence_compliance_finder.py module contains functions to find mentions of treatment adherence or compliance. The functions are named find_adherence_compliance_v1 through find_adherence_compliance_v5, where v1 is a high-recall version and v5 is a high-precision version.

Here's an example of how to use one of these functions:

from pyregularexpression.adherence_compliance_finder import find_adherence_compliance_v1

text = "The study measured adherence to the new drug. Adherence was defined as PDC > 0.8."

matches = find_adherence_compliance_v1(text)

for start, end, snippet in matches:
    print(f"Found '{snippet}' at indices {start}-{end}")

Example: Finding Medical Codes

Here is an example of how to use the medical_code_finder to extract potential medical codes from a piece of text.

from pyregularexpression.medical_code_finder import find_medical_code_v1

text = "The patient was diagnosed with ICD-10 code I21.0, which is an acute myocardial infarction. The CPT code was 99285."

matches = find_medical_code_v1(text)

for start, end, snippet in matches:
    print(f"Found medical code: {snippet}")

Helper Functions

The package also includes helper functions to apply multiple finder functions at once.

apply_regex_funcs

This function applies a list of finder functions to a text and returns a dictionary of the results.

from pyregularexpression.apply_regex_functions import apply_regex_funcs
from pyregularexpression.adherence_compliance_finder import find_adherence_compliance_v1
from pyregularexpression.algorithm_validation_finder import find_algorithm_validation_v1

text = "The study measured adherence to the new drug. The algorithm was validated."

results = apply_regex_funcs(text, [find_adherence_compliance_v1, find_algorithm_validation_v1])

print(results)

extract_regex_paragraphs_udf

This function returns a Spark UDF that can be used to extract paragraphs from a text that match any of a list of finder functions.

from pyspark.sql import SparkSession
from pyregularexpression.extract_regex_paragraphs_udf import extract_regex_paragraphs_udf
from pyregularexpression.adherence_compliance_finder import find_adherence_compliance_v1
from pyregularexpression.algorithm_validation_finder import find_algorithm_validation_v1

spark = SparkSession.builder.getOrCreate()

data = [("The study measured adherence to the new drug.\n\nThe algorithm was validated.",)]
df = spark.createDataFrame(data, ["text"])

udf = extract_regex_paragraphs_udf([find_adherence_compliance_v1, find_algorithm_validation_v1])

df.withColumn("matched_paragraphs", udf(df["text"])).show()