Description
IMPORTANT: Don’t run this model on the whole legal agreement. Instead:
- Split by paragraphs. You can use notebook 1 in Finance or Legal as inspiration;
- Use the
legclf_warranty_clauseText Classifier to select only these paragraphs;
This is a Legal Relation Extraction Model to identify the Subject (who), Action (web), Object(the indemnification) and Indirect Object (to whom) from warranty clauses.
Predicted Entities
is_warranty_indobject, is_warranty_object, is_warranty_subject
How to use
documentAssembler = nlp.DocumentAssembler()\
.setInputCol("text")\
.setOutputCol("document")
tokenizer = nlp.Tokenizer()\
.setInputCols("document")\
.setOutputCol("token")
embeddings = nlp.RoBertaEmbeddings.pretrained("roberta_embeddings_legal_roberta_base","en") \
.setInputCols(["document", "token"]) \
.setOutputCol("embeddings")
ner_model = legal.NerModel.pretrained('legner_warranty', 'en', 'legal/models') \
.setInputCols(["document", "token", "embeddings"]) \
.setOutputCol("ner")
ner_converter = nlp.NerConverter() \
.setInputCols(["document","token","ner"]) \
.setOutputCol("ner_chunk")
reDL = legal.RelationExtractionDLModel.pretrained("legre_warranty", "en", "legal/models") \
.setPredictionThreshold(0.5) \
.setInputCols(["ner_chunk", "document"]) \
.setOutputCol("relations")
pipeline = nlp.Pipeline(stages=[documentAssembler, tokenizer, embeddings, ner_model, ner_converter, reDL])
text = """ARTICLE XI - WARRANTIES 11.1 In addition to the warranties set forth in Article IX of the General Terms and Conditions of Transporter's FERC Gas Tariff, Shipper warrants the following: (a) Shipper warrants that all upstream and downstream transportation arrangements are in place, or will be in place as of the requested effective date of service, and that it has advised the upstream and downstream transporters of the receipt and delivery points under this Agreement and any quantity limitations for each point as specified on Exhibit "A" attached hereto."""
data = spark.createDataFrame([[text]]).toDF("text")
model = pipeline.fit(data)
res = model.transform(data)
Results
|relation |entity1 |entity1_begin|entity1_end|chunk1 |entity2 |entity2_begin|entity2_end|chunk2 |confidence|
|-------------------|----------------|-------------|-----------|--------|---------------|-------------|-----------|---------------------------------------------------------------------------------------------------------------------------------------|----------|
|is_warranty_subject|WARRANTY_SUBJECT|158 |164 |Shipper |WARRANTY_ACTION|166 |173 |warrants |0.98402506|
|is_warranty_subject|WARRANTY_SUBJECT|196 |202 |Shipper |WARRANTY_ACTION|204 |211 |warrants |0.9707028 |
|is_warranty_object |WARRANTY_SUBJECT|196 |202 |Shipper |WARRANTY |218 |352 |all upstream and downstream transportation arrangements are in place, or will be in place as of the requested effective date of service|0.9917001 |
|is_warranty_object |WARRANTY_SUBJECT|196 |202 |Shipper |WARRANTY |367 |474 |has advised the upstream and downstream transporters of the receipt and delivery points under this Agreement |0.79867786|
|is_warranty_object |WARRANTY_ACTION |204 |211 |warrants|WARRANTY |218 |352 |all upstream and downstream transportation arrangements are in place, or will be in place as of the requested effective date of service|0.97821265|
|is_warranty_object |WARRANTY_ACTION |204 |211 |warrants|WARRANTY |367 |474 |has advised the upstream and downstream transporters of the receipt and delivery points under this Agreement |0.80337876|
Model Information
| Model Name: | legre_warranty |
| Compatibility: | Legal NLP 1.0.0+ |
| License: | Licensed |
| Edition: | Official |
| Language: | en |
| Size: | 409.9 MB |
References
In-house annotated examples from CUAD legal dataset
Benchmarking
label Recall Precision F1 Support
is_warranty_indobject 1.000 1.000 1.000 15
is_warranty_object 1.000 1.000 1.000 44
is_warranty_subject 1.000 1.000 1.000 29
Avg 1.000 1.000 1.000 -
Weighted-Avg 1.000 1.000 1.000 -