Configuration values & examples
A list of all settable values associated with a Semantria configuration
The mandatory values for creating a configuration are:
name
language_id
During config creation, an ID is assigned by Semantria. During update or delete, you pass the ID of the configuration you want to manipulate to Semantria in the URL.
General Settings
Setting | Default Value | Type | Description |
---|---|---|---|
|
| integer | Percentage of content that has to be alpha-numeric for processing to succeed. Documents that do not meet this threshold will be returned to you with a FAILED status. |
|
| string | URL Semantria should POST data to when processed. Only set this if you are using the callback data retrieval method. If this is set, Semantria will enable callback and send documents to this URL. The URL must start with 'https://' |
|
| double | Confidence level a category (concept topic) match must have to be returned for a document. |
|
| double | Confidence level an entity match must have to be returned for a document. |
|
| boolean | Whether to use the auto-response method for doc processing. |
| no default | string | Name of configuration. |
| no default | string | The two character ID of the language you are going to process with this config. |
| no default | string | Which language you are going to process with this config. This will be auto-filled by Semantria from the code when first processing. |
|
| string | Which Parts of Speech to return |
|
| integer | Length of summary in sentences to return per document. |
| no default | string | Template id this config was created from, if any. |
| no default |
| UTC date and time this config was last updated. |
Boolean Settings
The presence of an argument in the below settings turns that setting on, while absence turns the setting off
| Option to turn on/off anaphora resolution. |
---|---|
| Turn off/on attribute calculation for collection mode. |
| Return automatic classification based on a pre-built taxonomy (not editable by the user). |
| Option to respect lists of words that should not be returned as part of sentiment or themes. |
| Return Concept Topics. |
| Use Semantria-provided language detection library. |
| Use Semantria algorithm to detect possible sentiment phrases. |
| Use user-provided model-based document classification. |
| Stem entities. |
| Can a single string return two entities that overlap. E.g. "Ford Fusion" is both a car and a company. |
| Calculate facets in collection mode. |
| Stem facets. |
| Fail documents with a sentence longer than 1,000 words. |
| Turn all upper-case documents into lower-case for NLP purposes. |
| Return intentions. |
| Return individual mention location information for NLP extracted items such as entities, themes and queries. |
| Run user-provided model based sentiment. |
| Return entities discovered by NLP. |
| Treat entire document as a single sentence for NLP purposes. |
| Return opinions. |
| Don't use built-in sentiment dictionary, only user-provided one. |
| Try to extract text from well-formed HTML. This will treat the document as having HTML markup. |
| Return Query Topics. |
| Return themes. |
| Whether to retrieve intentions for a document or not. |
| Return sentiment phrases. |
| Return relations. |
| Return extracted sentiment phrases. |
| Return results from user-defined taxonomy. |
| Stem content for query topics. |
| Return user-defined entities. |
Example JSON output
An example of a JSON object for setting the values is below.
{
"alphanumeric_threshold": 41,
"concept_topics_threshold": 0.35,
"entities_threshold": 40,
"id": "ed7b6405-2bc2-443d-b6c4-0feab9050c5d",
"is_autoresponse_enabled": false,
"language_id": "en",
"language_name": "English",
"name": "english --- default",
"pos_types": "all",
"settings": [
"OPINIONS",
"FACETS",
"MENTIONS",
"NAMED_ENTITIES",
"SENTIMENT_PHRASES",
"TAXONOMY",
"ATTRIBUTES",
"ENTITIES_STEMMING",
"AUTO_CATEGORIES",
"ENTITY_OVERLAP",
"QUERY_TOPICS",
"RELATIONS",
"FAIL_ON_LONG_SENTENCES",
"FACETS_STEMMING",
"DOC_CLASSIFICATION",
"ONE_SENTENCE_MODE",
"DETECT_LANGUAGE",
"FLATTEN_ALL_UPPER_CASE",
"TOPIC_STEMMING",
"USER_ENTITIES",
"RETURN_SOURCE_TEXT",
"MODEL_SENTIMENT",
"INTENTIONS",
"OVERRIDE_SENTIMENT_DICTIONARY",
"DETECT_PHRASES",
"ANAPHORA_RESOLUTION",
"BLACKLISTS",
"CONCEPT_TOPICS",
"THEMES",
"PROCESS_HTML"
],
"summary_size": 3,
"updated": "2021-02-18T20:46:22.927Z"
}
Updated about 1 year ago