Excerpt |
---|
Looks up matches to regular expressions in a dictionary across mutiple multiple tokens and then tags the match with one or more semantic tags as an alternative representation(s). For a simple regex expression that , a match only needs to match occur against a singe token. Simple Regex Stage is recomendedrecommended. |
Operates On: Lexical Items with TOKEN flag
Note |
---|
All possibilities are tagged , (including overlaps and sub-patterns, ) with the expectation that later disambiguation stages will choose which tags are the correct interpretation. |
Warning |
---|
This stage requires a lot of process processing time, please . Please follow this these recommendations: - keep Keep the amount at a minimum the amount to regex patterns.
- try Try to use non greedy regex.
- set Set the maxLength to the bare minimum necessary for the expected matches.
|
Include Page |
---|
| Generic Configuration Parameters |
---|
| Generic Configuration Parameters |
---|
|
...
- patterns (string, required) - The resource which that contains the pattern database.
- See below for the format.
- maxLength (integer, optional) - The max length of text to test for regex, default . The default is 25 characters.
- For each token, the stage will increase the size by adding tokens before and after, until a match (or the 25 character limit) is reachreached.
- caseInsensitive (boolean, optional) - If true, all regex will be process as case insensitive (default = true)
...
In the following example, "What's your name" is in the dictionary as a regex for self-name, and there are also regex for number numbers "[0-9]+" and "[0-9]+\\.[0-9]+" :
Code Block |
---|
language | text |
---|
theme | FadeToGrey |
---|
|
V--------------------------------------[What's your name 12 @#$ 25 63.3]---------------------------------------V
^-----[What's]-----V--[your]--V--[name]--V-----[12]-----V--[@#$]--V-----[25]-----V-----------[63.3]------------^
^--[What]--V--[s]--^ ^--[{number}]--^ ^--[{number}]--^-----[63]-----V-----[3]------^
^-----[what's]-----^ ^---------[{number}]----------^
^--[what]--^ ^--[{number}]--^--[{number}]--^
^-------------[{self-name}]--------------^
|
Output Flags
Lex-Item Flags
...
- SEMANTIC_TAG - Identifies all lexical items which that are semantic tags.
- PROCESSED - Placed on all the tokens which composed composing the semantic tag.
Resource Data
The regex pattern must have an a "pattern dictionary" (a string to JSON map) which is a list of JSON records, indexed by entity ID. In addition, there may also be a pattern map and a token index.
Pattern (Regex) Dictionary Format
The only required file which is absolutely required is the pattern dictionary. It is a series of JSON records, typically indexed by entity ID.
...
- Multiple patterns can have the same entry.
- Additional fielded data can be added to the record.
- As needed by downstream processes.
Fields
- id (required, string) - Identifies the entity by unique ID. This identifier must be unique across all entries (across all dictionaries).
- Typically, this is an identifier with has meaning to the larger application which that is using the Language Processing Toolkit.
- tags (required, array of string) - The list of semantic tags which will be added to add to the interpretation graph whenever any of the patterns are matched.
- These will all be added to the interpretation graph with the SEMANTIC_TAG flag.
- patterns (required, array of string) - A list of patterns to match in the content.
- splitMatch (optional, boolean) - Indicates if whether or not the partialmatch will create a regex tag even if a full match was not met.
- confidence (optional, float) - Specifies the confidence level of the entity, independent of any patterns matched.
- This is the confidence of the entry, in comparison to all of the other entries. Essentially, the likelihood that this entry , will be randomly encountered.
...
- display (optional, string) - What to show the user when browsing this the entity.
- context (optional, object) - A context vector which that can help disambiguate this the entity from others with the same pattern.
- Format TBD, but probably a list of weighted words, phrases and tags.
...