AWS Comprehend Client Package

aws.comprehend is a package for natural language processing.

Code Examples

All of the functions (except detect_medical_*) accept either a single character string or a character vector. Note that AWS currently limits batch queries to 25 documents, so character vectors should have 25 elements maximum.

The default language is English ("en") but this is easily changed using the language argument.

Sentiment analysis

library("aws.comprehend")

detect_sentiment("I have never been happier. This is the best day ever.")

##   Index Sentiment       Mixed     Negative      Neutral  Positive
## 1     0  POSITIVE 1.21042e-06 5.316024e-05 0.0003428663 0.9996029

# Sentiment analysis in Spanish
detect_sentiment("¡Hoy estoy feliz!", language = "es")

##   Index Sentiment        Mixed    Negative    Neutral  Positive
## 1     0  POSITIVE 0.0001126147 0.002433205 0.03607949 0.9613748

Language detection

# simple language detection
detect_language("This is a test sentence in English")

##   Index LanguageCode     Score
## 1     0           en 0.9729235

# multi-lingual language detection
detect_language("A: ¡Hola! ¿Como está, usted? B: Bien, merci. Et toi?")

##   Index LanguageCode     Score
## 1     0           fr 0.7126021
## 2     0           es 0.2452095

Named Entity Recognition

txt <- c("Amazon provides web services.", "Jeff is their leader.")
detect_entities(txt)

##   Index BeginOffset EndOffset     Score   Text         Type
## 1     0           0         6 0.9992782 Amazon ORGANIZATION
## 2     1           0         4 0.9999498   Jeff       PERSON

Key Phrase Detection

txt <- c("Amazon provides web services.", "Jeff is their leader.")
detect_phrases(txt)

##   Index BeginOffset EndOffset Score         Text
## 1     0           0         6     1       Amazon
## 2     0          16        28     1 web services
## 3     1           0         4     1         Jeff
## 4     1           8        20     1 their leader

Syntax Analysis

detect_syntax("The quick fox jumps over the lazy dog.")

##   Index BeginOffset EndOffset PartOfSpeech.Score PartOfSpeech.Tag  Text TokenId
## 1     0           0         3          0.9999670              DET   The       1
## 2     0           4         9          0.9966556              ADJ quick       2
## 3     0          10        13          0.9957780             NOUN   fox       3
## 4     0          14        19          0.8895551             VERB jumps       4
## 5     0          20        24          0.9910401              ADP  over       5
## 6     0          25        28          0.9999968              DET   the       6
## 7     0          29        33          0.9885939              ADJ  lazy       7
## 8     0          34        37          0.9999415             NOUN   dog       8
## 9     0          37        38          0.9999982            PUNCT     .       9

Medical Entity and Personal Health Information (PHI) Detection

# medical entity detection
medical_txt <- "Pt is 40yo mother, highschool teacher. HPI : Sleeping trouble on present dosage of Clonidine."
detect_medical_entities(medical_txt)

##   Index BeginOffset                     Category EndOffset Id     Score               Text                    Traits         Type
## 1     0           6 PROTECTED_HEALTH_INFORMATION        10  2 0.9982511               40yo                      NULL          AGE
## 2     0          19 PROTECTED_HEALTH_INFORMATION        37  3 0.4113526 highschool teacher                      NULL   PROFESSION
## 3     0          45            MEDICAL_CONDITION        61  1 0.7587468   Sleeping trouble SYMPTOM, 0.52603405714035      DX_NAME
## 4     0          83                   MEDICATION        92  0 0.9932888          Clonidine                      NULL GENERIC_NAME

# Protected Health Information (PHI) detection
detect_medical_phi(medical_txt)

##   Index BeginOffset                     Category EndOffset Id     Score               Text Traits       Type
## 1     0           6 PROTECTED_HEALTH_INFORMATION        10  0 0.9982511               40yo   NULL        AGE
## 2     0          19 PROTECTED_HEALTH_INFORMATION        37  1 0.4113526 highschool teacher   NULL PROFESSION

Setting up credentials

To use the package, you will need an AWS account and to enter your credentials into R. Your keypair can be generated on the IAM Management Console under the heading Access Keys. Note that you only have access to your secret key once. After it is generated, you need to save it in a secure location. New keypairs can be generated at any time if yours has been lost, stolen, or forgotten. The aws.iam package profiles tools for working with IAM, including creating roles, users, groups, and credentials programmatically; it is not needed to use IAM credentials.

A detailed description of how credentials can be specified is provided at: https://github.com/cloudyr/aws.signature/. The easiest way is to simply set environment variables on the command line prior to starting R or via an Renviron.site or .Renviron file, which are used to set environment variables in R during startup (see ? Startup). They can be also set within R:

Sys.setenv("AWS_ACCESS_KEY_ID" = "mykey",
           "AWS_SECRET_ACCESS_KEY" = "mysecretkey",
           "AWS_DEFAULT_REGION" = "us-east-1",
           "AWS_SESSION_TOKEN" = "mytoken")

Installation

You can install this package from CRAN or, to install the latest development version, from the cloudyr drat repository:

# Install from CRAN
install.packages("aws.comprehend")

# Latest version passing CI tests, from drat repo
install.packages("aws.comprehend", repos = c(getOption("repos"), "http://cloudyr.github.io/drat"))

You can also pull a potentially unstable version directly from GitHub, using the remotes package:

remotes::install_github("cloudyr/aws.comprehend")

---