This API provides numerous text analysis features. There are several beta features which are not documented.
The analysis from this API is designed to be functionally similar to those provided by IBM Watson, but because of differences in the scoring, and what we feel is a greater accuracy in this API, the results may not be 100% compatible.
This API has also been optimized for ease of use. You do not need several API's to achieve most text analysis tasks.
Because you can have multiple analysis in a single call, by using comma separated values, most calls will have multiple analysis performed, and if the same text is to be analyzed several ways, this will save you in time, and cost.
You can run analysis by adding them to the "m" portion of the request (m = method)
Available methods include:
** Coming soon:**
Descriptions of Methods:
partofspeechbysent - Part of speech for each word grouped by sentence.
nounentitiesbysent - Noun Entities/Named Entities per sentence.
flagsbysent - Flags are signals about the sentence. Flags which start with a number 1-6 are final. Flags with a . are beta. Flags with no . are alpha.
tokencount - Number of tokens in the text.
interogativebysent - The who, what, where, when, why, and how of the sentences.
questiontypebysent - The type of question being asked by sentence. "How big is a blue whale" is a question of dimension as an example. There are currently 250 types.
intent - A description of what the user intends to have happen [Beta].
facts - The facts contained in the sentences of the text.
opedness - The ammount of opinion conveyed in the text.
strunk - Edits necessary to conform with the Strunk and White Style guide.
alltokennosuffix - Suffixes removed from tokens such that dumb and dumber are the same token.
tokencounts - Counts of all tokens in the text.
lctokencounts - Lower Case token counts.
nounentitycounts - Counts of named/noun entities.
flagcounts - Counts of flags.
summarytext - 350 character abstract of the text.
composition - the make up of the text
synonymcomposition - Composition based on synonyms or concepts of the text.
traits - Traits of the text.
index - Index of the words by sentence which contains them.
scoredsentences - Importance of each sentence. Can be used for summarization.
lcngramsbysent - nGrams contained in each sentence.
lcngrams - All lower case nGrams for the text.
ngramsbysent - Case maintined nGrams by sentence.
ngrams - Case maintined nGrams.
wordbag - All of the words in the text.
lcwordbag - All of the words as lower case in the text.
sylcountbyword - Syllable counts by word.
sylcountbysent - Syllable counts by sentence.
sylcounttotal - Total syllable count.
prominencebyword - Prominence of words (how rare they are).
prominenceaverage - Average prominence.
charfrequency - Character frequency (useful for determining if English, or if Lyric)
aoaaverage - Average age of Acquisition.
reducedtokens - High collision rate tokenization method, useful for duplicate content detection or broad search.
reducednumeric - Reduced significant digit numbers to aid in collisions for search, or fact validation.
contenthash - Hash of the content for aiding in collision detection.
tense - The tense of each sentence in the text.
phoneticsbysent - Finding collisions of names often requires multiple possible phonetic matches. This returns all possible pronunciations for each word in each sentence.
sentiment - The sentiment of each sentence. Negative values indicate negative sentiment.
lolspeak - Translates the text to Lolspeak. We iz not suah whut diz iz useful for, but eet iz fun.