Medical Record Contextual Parser Pipeline

Description

This pipeline, extracts medical_record entities from clinical texts.

Copy S3 URI

How to use


from sparknlp.pretrained import PretrainedPipeline

pipeline = PretrainedPipeline("med_parser_pipeline", "en", "clinical/models")

sample_text = """ Month DD, YYYY
XYZ
RE: ABC
MEDICAL RECORD#: 12332
MRN: 1233567
MEDICAL RECORD#: 45678
MRN: 9876543
MEDICAL RECORD#: 11111
Dear Dr. XYZ:

I saw ABC back in Neuro-Oncology Clinic today."""

result = pipeline.transform(spark.createDataFrame([[sample_text]]).toDF("text"))


from johnsnowlabs import nlp, medical

pipeline = nlp.PretrainedPipeline("med_parser_pipeline", "en", "clinical/models")

sample_text = """ Month DD, YYYY
XYZ
RE: ABC
MEDICAL RECORD#: 12332
MRN: 1233567
MEDICAL RECORD#: 45678
MRN: 9876543
MEDICAL RECORD#: 11111
Dear Dr. XYZ:

I saw ABC back in Neuro-Oncology Clinic today."""

result = pipeline.transform(spark.createDataFrame([[sample_text]]).toDF("text"))


import com.johnsnowlabs.nlp.pretrained.PretrainedPipeline

val pipeline = PretrainedPipeline("med_parser_pipeline", "en", "clinical/models")

val sample_text = """ Month DD, YYYY
XYZ
RE: ABC
MEDICAL RECORD#: 12332
MRN: 1233567
MEDICAL RECORD#: 45678
MRN: 9876543
MEDICAL RECORD#: 11111
Dear Dr. XYZ:

I saw ABC back in Neuro-Oncology Clinic today."""

val result = pipeline.transform(spark.createDataFrame([[sample_text]]).toDF("text"))

Results


|   chunk |   begin |   end | label         |
|--------:|--------:|------:|:--------------|
|   12332 |      44 |    48 | MEDICALRECORD |
| 1233567 |      55 |    61 | MEDICALRECORD |
|   45678 |      80 |    84 | MEDICALRECORD |
| 9876543 |      91 |    97 | MEDICALRECORD |
|   11111 |     116 |   120 | MEDICALRECORD |

Model Information

Model Name: med_parser_pipeline
Type: pipeline
Compatibility: Healthcare NLP 6.3.0+
License: Licensed
Edition: Official
Language: en
Size: 396.6 KB

Included Models

  • DocumentAssembler
  • SentenceDetectorDLModel
  • TokenizerModel
  • ContextualParserModel
  • ChunkConverter