Selective deep parsing of natural language content

1. A method comprising:

configuring a targeted deep parse natural language processing (NLP) system to recognize a plurality of triggers, wherein the triggers are elements within natural language content that indicate content that is to be targeted with a deep parse operation, wherein the targeted deep parse NLP system is configured, for each trigger, with a corresponding granularity of a corresponding targeted sub-portion of the natural language content to perform a deep parse in response to detecting the trigger in the natural language content, and wherein at least two triggers in the plurality of triggers have different granularities;

performing, by the targeted deep parse NLP system, a pre-deep parse scan operation of a portion of natural language content, from a corpus of natural language content, based on the plurality of triggers to identify one or more sub-portions of the portion of natural language content that contain at least one of the triggers; and

performing, by the targeted deep parse NLP system, a NLP operation comprising a deep parse of only the one or more sub-portions of the portion of natural language content that contain at least one of the triggers, while other sub-portions of the portion of natural language content are not deep parsed, wherein the targeted deep parse NLP system is configured with configuration information associated with a plurality of annotators, selected from a library of annotators, that are enabled in a cognitive computing system downstream of the targeted deep parse NLP system, wherein the configuration information specifies triggers, in the plurality of triggers, for each of the plurality of annotators, and wherein at least two of the annotators in the plurality of annotators have different associated triggers.

2. The method of claim 1 , wherein configuring the targeted deep parse NLP system comprises determining which annotators, in the library of annotators are enabled in the cognitive computing system to process the portion of natural language content, to form determined annotators, wherein the configuration information associated with the plurality of annotators is configuration information associated with the determined annotators.

3. The method of claim 1 , wherein performing the pre-deep parse scan operation of the portion of natural language content comprises associating with each of the one or more sub-portions of the portion of natural language content, a deep parse indicator specifying that the one or more sub-portions of the portion of natural language content are to be deep parsed by a deep parser of the targeted deep parse NLP system.

4. The method of claim 3 , wherein performing the NLP operation comprises executing a deep parser on the portion of natural language content, wherein the deep parser skips sub-portions of the portion of natural language content that do not have associated deep parse indicators.

5. The method of claim 1 , wherein performing the NLP operation comprises generating a parse tree for each of the one or more sub-portions and performing a NLP operation on the parse trees for the one or more sub-portions.

6. The method of claim 1 , wherein the one or more sub-portions of the portion of natural language content comprises at least one portion of metadata associated with the natural language content.

7. The method of claim 1 , wherein the granularity specified in the configuration information comprises, for at least one of the triggers in the plurality of triggers, an exclusion granularity specifying a sub-portion of the portion of natural language content that is to be excluded from a deep parse by a deep parser as part of the performance of the natural language processing operation.

8. The method of claim 1 , wherein each annotator, in the library of annotators, is configured to extract at least one of concepts, words, phrases, classifications, or named entities, for a corresponding domain, from unstructured natural language content to form extractions and annotate the unstructured natural language content based on the extractions.

9. The method of claim 8 , wherein the plurality of annotators are selected from the library of annotators for enabling in the cognitive computing system based on an operation to be performed by the cognitive computing system and a type of the annotators in the plurality of annotators.

10. The method of claim 1 , wherein at least one annotator in the plurality of annotators, is a negation term annotator that annotates unstructured natural language content that comprises negation terms.

11. A computer program product comprising a computer readable storage medium having a computer readable program stored therein, wherein the computer readable program, when executed on a computing device, causes the computing device to:

configure a targeted deep parse natural language processing (NLP) system to recognize a plurality of triggers, wherein the triggers are elements within natural language content that indicate content that is to be targeted with a deep parse operation, wherein the targeted deep parse NLP system is configured, for each trigger, with a corresponding granularity of a corresponding targeted sub-portion of the natural language content to perform a deep parse in response to detecting the trigger in the natural language content, and wherein at least two triggers in the plurality of triggers have different granularities;

perform, by the targeted deep parse NLP system, a pre-deep parse scan operation of a portion of natural language content, from a corpus of natural language content, based on the plurality of triggers to identify one or more sub-portions of the portion of natural language content that contain at least one of the one or more triggers; and

perform, by the targeted deep parse NLP system, a NLP operation comprising a deep parse of only the one or more sub-portions of the portion of natural language content that contain at least one of the triggers, while other sub-portions of the portion of natural language content are not deep parsed, wherein the targeted deep parse NLP system is configured with configuration information associated with a plurality of annotators, selected from a library of annotators, that are enabled in a cognitive computing system downstream of the targeted deep parse NLP system, wherein the configuration information specifies triggers, in the plurality of triggers, for each of the plurality of annotators, and wherein at least two of the annotators in the plurality of annotators have different associated triggers.

12. The computer program product of claim 11 , wherein configuring the targeted deep parse NLP system comprises determining which annotators, in the library of annotators are enabled in the cognitive computing system to process the portion of natural language content, to form determined annotators, wherein the configuration information associated with the plurality of annotators is configuration information associated with the determined annotators.

13. The computer program product of claim 11 , wherein performing the pre-deep parse scan operation of the portion of natural language content comprises associating with each of the one or more sub-portions of the portion of natural language content, a deep parse indicator specifying that the one or more sub-portions of the portion of natural language content are to be deep parsed by a deep parser of the targeted deep parse NLP system.

14. The computer program product of claim 13 , wherein performing the NLP operation comprises executing a deep parser on the portion of natural language content, wherein the deep parser skips sub-portions of the portion of natural language content that do not have associated deep parse indicators.

15. The computer program product of claim 11 , wherein performing the NLP operation comprises generating a parse tree for each of the one or more sub-portions and performing a NLP operation on the parse trees for the one or more sub-portions.

16. The computer program product of claim 11 , wherein the one or more sub-portions of the portion of natural language content comprises at least one portion of metadata associated with the natural language content.

17. The computer program product of claim 11 , wherein the granularity specified in the configuration information comprises, for at least one of the triggers in the plurality of triggers, an exclusion granularity specifying a sub-portion of the portion of natural language content that is to be excluded from a deep parse by a deep parser as part of the performance of the natural language processing operation.

18. The computer program product of claim 11 , wherein each annotator, in the library of annotators, is configured to extract at least one of concepts, words, phrases, classifications, or named entities, for a corresponding domain, from unstructured natural language content to form extractions and annotate the unstructured natural language content based on the extractions.

19. The computer program product of claim 18 , wherein the plurality of annotators are selected from the library of annotators for enabling in the cognitive computing system based on an operation to be performed by the cognitive computing system and a type of the annotators in the plurality of annotators.

20. An apparatus comprising:

a processor; and

a memory coupled to the processor, wherein the memory comprises instructions which, when executed by the processor, cause the processor to:

configure a targeted deep parse natural language processing (NLP) system to recognize a plurality of triggers, wherein the triggers are elements within natural language content that indicate content that is to be targeted with a deep parse operation, wherein the targeted deep parse NLP system is configured, for each trigger, with a corresponding granularity of a corresponding targeted sub-portion of the natural language content to perform a deep parse in response to detecting the trigger in the natural language content, and wherein at least two triggers in the plurality of triggers have different granularities;

perform, by the targeted deep parse NLP system, a pre-deep parse scan operation of a portion of natural language content, from a corpus of natural language content, based on the plurality of triggers to identify one or more sub-portions of the portion of natural language content that contain at least one of the one or more triggers; and

perform, by the targeted deep parse NLP system, a NLP operation comprising a deep parse of only the one or more sub-portions of the portion of natural language content that contain at least one of the triggers, while other sub-portions of the portion of natural language content are not deep parsed, wherein the targeted deep parse NLP system is configured with configuration information associated with a plurality of annotators, selected from a library of annotators, that are enabled in a cognitive computing system downstream of the targeted deep parse NLP system, wherein the configuration information specifies triggers, in the plurality of triggers, for each of the plurality of annotators, and wherein at least two of the annotators in the plurality of annotators have different associated triggers.