Sign up for the Healthcare NLP Summit 2024, April 2-3. Register now.
was successfully added to your cart.

1 line of code to visualize dependency trees, entity relationships, recognition, resolution, and assertion in NLU 3

NLU 3 is the most visually appealing release to date, with the integration of the Spark-NLP-Display library and visualizations for dependency trees, entity resolution, entity assertion, the relationship between entities, and named entity recognition. In addition to this, the schema of how columns are named by NLU has been reworked and all 140+ tutorial notebooks have been updated to reflect the latest changes in NLU 3.

Finally, new multilingual models for Afrikaans, Welsh, Maltese, Tamil, and Vietnamese are now available.

New Features and Enhancements

  • 1 line to visualization for NER, Dependency, Resolution, Assertion, and Relation via Spark-NLP-Display integration
  • Improved column naming schema
  • Over 140 + NLU tutorial Notebooks updated and improved to reflect the latest changes in NLU 3.0.0 +
  • New multilingual models for Afrikaans, Welsh, Maltese, Tamil, and Vietnamese
  • Enhanced offline loading

NLU visualization

The latest NLU release integrated the beautiful Spark-NLP-Display package visualizations. You do not need to worry about installing it, when you try to visualize something, NLU will check if
Spark-NLP-Display is installed, if it is missing it will be dynamically installed into your python executable environment, so you don’t need to worry about anything!

See the visualization tutorial notebook and visualization docs for more info.

Cheat Sheet visualization

NER visualization

Applicable to any of the 100+ NER models! See here for an overview

nlu.load('ner').viz("Donald Trump from America and Angela Merkel from Germany don't share many oppinions.")
NER visualization

Dependency tree visualization

Visualizes the structure of the labeled dependency tree and part of speech tags

nlu.load('dep.typed').viz("Billy went to the mall")
Dependency Tree visualization
#Bigger Example
nlu.load('dep.typed').viz("Donald Trump from America and Angela Merkel from Germany don't share many oppinions but they both love John Snow Labs software")
Dependency Tree visualization

Assertion status visualization

Visualizes asserted statuses and entities.
Applicable to any of the 10 + Assertion models! See here for an overview

nlu.load('med_ner.clinical assert').viz("The MRI scan showed no signs of cancer in the left lung")
Assert visualization
#bigger example
data ='This is the case of a very pleasant 46-year-old Caucasian female, seen in clinic on 12/11/07 during which time MRI of the left shoulder showed no evidence of rotator cuff tear. She did have a previous MRI of the cervical spine that did show an osteophyte on the left C6-C7 level. Based on this, negative MRI of the shoulder, the patient was recommended to have anterior cervical discectomy with anterior interbody fusion at C6-C7 level. Operation, expected outcome, risks, and benefits were discussed with her. Risks include, but not exclusive of bleeding and infection, bleeding could be soft tissue bleeding, which may compromise airway and may result in return to the operating room emergently for evacuation of said hematoma. There is also the possibility of bleeding into the epidural space, which can compress the spinal cord and result in weakness and numbness of all four extremities as well as impairment of bowel and bladder function. However, the patient may develop deeper-seated infection, which may require return to the operating room. Should the infection be in the area of the spinal instrumentation, this will cause a dilemma since there might be a need to remove the spinal instrumentation and/or allograft. There is also the possibility of potential injury to the esophageus, the trachea, and the carotid artery. There is also the risks of stroke on the right cerebral circulation should an undiagnosed plaque be propelled from the right carotid. She understood all of these risks and agreed to have the procedure performed.'
nlu.load('med_ner.clinical assert').viz(data)
Assert visualization

Relationship between entities visualization

Visualizes the extracted entities between relationships. Applicable to any of the 20 + Relation Extractor models See here for an overview

nlu.load('med_ner.jsl.wip.clinical relation.temporal_events').viz('The patient developed cancer after a mercury poisoning in 1999 ')
Entity Relation visualization
# bigger example
data = 'This is the case of a very pleasant 46-year-old Caucasian female, seen in clinic on 12/11/07 during which time MRI of the left shoulder showed no evidence of rotator cuff tear. She did have a previous MRI of the cervical spine that did show an osteophyte on the left C6-C7 level. Based on this, negative MRI of the shoulder, the patient was recommended to have anterior cervical discectomy with anterior interbody fusion at C6-C7 level. Operation, expected outcome, risks, and benefits were discussed with her. Risks include, but not exclusive of bleeding and infection, bleeding could be soft tissue bleeding, which may compromise airway and may result in return to the operating room emergently for evacuation of said hematoma. There is also the possibility of bleeding into the epidural space, which can compress the spinal cord and result in weakness and numbness of all four extremities as well as impairment of bowel and bladder function. However, the patient may develop deeper-seated infection, which may require return to the operating room. Should the infection be in the area of the spinal instrumentation, this will cause a dilemma since there might be a need to remove the spinal instrumentation and/or allograft. There is also the possibility of potential injury to the esophageus, the trachea, and the carotid artery. There is also the risks of stroke on the right cerebral circulation should an undiagnosed plaque be propelled from the right carotid. She understood all of these risks and agreed to have the procedure performed'
pipe = nlu.load('med_ner.jsl.wip.clinical relation.clinical').viz(data)
Entity Relation visualization

Entity Resolution visualization for chunks

Visualizes resolutions of entities
Applicable to any of the 100+ Resolver models See here for an overview

nlu.load('med_ner.jsl.wip.clinical resolve_chunk.rxnorm.in').viz("He took Prevacid 30 mg  daily")
Chunk Resolution visualization
# bigger example
data = "This is an 82 - year-old male with a history of prior tobacco use , hypertension , chronic renal insufficiency , COPD , gastritis , and TIA who initially presented to Braintree with a non-ST elevation MI and Guaiac positive stools , transferred to St . Margaret\'s Center for Women & Infants for cardiac catheterization with PTCA to mid LAD lesion complicated by hypotension and bradycardia requiring Atropine , IV fluids and transient dopamine possibly secondary to vagal reaction , subsequently transferred to CCU for close monitoring , hemodynamically stable at the time of admission to the CCU ."
nlu.load('med_ner.jsl.wip.clinical resolve_chunk.rxnorm.in').viz(data)
Chunk Resolution visualization

Entity Resolution visualization for sentences

Visualizes resolutions of entities in sentences
Applicable to any of the 100+ Resolver models See here for an overview

nlu.load('med_ner.jsl.wip.clinical resolve.icd10cm').viz('She was diagnosed with a respiratory congestion')
Sentence Resolution visualization
# bigger example
data = 'The patient is a 5-month-old infant who presented initially on Monday with a cold, cough, and runny nose for 2 days. Mom states she had no fever. Her appetite was good but she was spitting up a lot. She had no difficulty breathing and her cough was described as dry and hacky. At that time, physical exam showed a right TM, which was red. Left TM was okay. She was fairly congested but looked happy and playful. She was started on Amoxil and Aldex and we told to recheck in 2 weeks to recheck her ear. Mom returned to clinic again today because she got much worse overnight. She was having difficulty breathing. She was much more congested and her appetite had decreased significantly today. She also spiked a temperature yesterday of 102.6 and always having trouble sleeping secondary to congestion'
nlu.load('med_ner.jsl.wip.clinical resolve.icd10cm').viz(data)
Sentence Resolution visualization

Configure visualizations

Define custom colors for labels

Some entity and relation labels will be highlighted with a pre-defined color, which you can find here.
For labels that have no color defined, a random color will be generated.
You can define colors for labels manually, by specifying via the viz_colors parameter and defining hex color codes in a dictionary that maps labels to colors.

data = 'Dr. John Snow suggested that Fritz takes 5mg penicilin for his cough'
# Define custom colors for labels
viz_colors={'STRENGTH':'#800080', 'DRUG_BRANDNAME':'#77b5fe', 'GENDER':'#77ffe'}
nlu.load('med_ner.jsl.wip.clinical').viz(data,viz_colors =viz_colors)
define colors labels

Filter entities that get highlighted

By default, every entity class will be visualized.
The labels_to_viz can be used to define a set of labels to highlight. Applicable for ner, resolution, and assert.

data = 'Dr. John Snow suggested that Fritz takes 5mg penicilin for his cough'
# Filter wich NER label to viz
labels_to_viz=['SYMPTOM']
nlu.load('med_ner.jsl.wip.clinical').viz(data,labels_to_viz=labels_to_viz)
filter labels

Reworked and updated NLU tutorial notebooks

All of the 140+ NLU tutorial Notebooks have been updated and reworked to reflect the latest changes in NLU 3

Column Name generation in NLU 3

  • NLU categorized each internal component now with boolean labels for name_deductable and always_name_deductable .
  • Before generating column names, NLU checks whether each component is unique in the pipeline or not. If a component is not unique in the
    pipe and there are multiple components of the same type, i.e. multiple NER models, NLU will deduct a base name for the final output columns from the NLU reference each NER model is pointing to.
  • If on the other hand, there is only one NER model in the pipeline, only the default ner column prefixed will be generated.
  • For some components, like embeddings and classifiers are now defined as always_name_deductable, for those NLU will always try to infer a meaningful base name for the output columns.
  • Newly trained component output columns will now be prefixed with trained_<type> , for types pos, ner, cLassifier, sentiment and multi_classifier

Enhanced offline mode in NLU 3

  • You can still load a model from a path as usual with nlu.load(path=model_path) and output columns will be suffixed with from_disk
  • You can now optionally also specify request parameter during load a model from HDD, it will be used to deduct more meaningful column name suffixes, instead of from_disk, i.e. by calling nlu.load(request =’en.embed_sentence.biobert.pubmed_pmc_base_cased’, path=model_path)

Additional NLU resources

1 line Install NLU on Google Colab

!wget https://setup.johnsnowlabs.com/nlu/colab.sh -O - | bash

1 line Install NLU on Kaggle

!wget https://setup.johnsnowlabs.com/nlu/kaggle.sh -O - | bash

Install via PIP

! pip install nlu pyspark==3.0.1

1 Line of Code to Train A Multilingual Text Classifier for 100+ Languages with NLU 1.1.4

We are very excited to announce NLU 1.1.4 has been released and comes with a lot of tutorials showcasing how you can...
preloader