Data Type

Specifies the base type used for interpreting captured data values. If a captured value cannot be converted to the Value Type, it will be excluded from the output by default. The 'Allow Invalid Results' property of the Result Filter can be used to override this behavior. If the specified type is a formattable type, each output value will be reformatted based an a configurable Format Specifier.

An optional list of cultures supported by this extractor. If this value is empty, the extractor will execute against all documents. Otherwise, the extractor will only execute on documents with a compatible culture code.

Specifies a description for the item.

Defines an internal Data Pattern which can be used in place of a child Data Format. This property is useful for simple extractions where only a single format needs to be defined.

Defines an optional list of external extractors to be executed. At runtime, referenced extractors execute after the internal pattern and the direct children have been executed.

An optional extractor which restricts the scope of data extraction to a subset of the input. When an Input Filter is specified, extraction logic executes against each match produced by the Input Filter, rather than against the input itself. This mechanism can be used to restrict extraction operations to a subset of the original input.

In some cases, input filters are used simplify data extraction logic by restricting the scope to a specific page or section of the document. In other cases, input filters are used to reduce the execution time of fuzzy extractors by limiting the amount of content they execute against.

The following examples demonstrate some commonly-used input filters:

• Restrict to the first page of the document: ^[^f]+
• Restrict to the last page of the document: [^f]+$
• Restrict to first 5 lines of the document: ^([^\r\n]+\r\n){5}
• Restrict to top 3 lines of each page: (^|\f)([^\r\n]+\r\n){3}
• Restrict to the PERSONAL INFO section: \r\nPERSONAL INFO[^\f]+\r\nEDUCATION

An optional extractor to be used for filtering undesirable results from the result set. Any result which overlaps with an exclusion match will be discarded.

An optional extractor to be used for removing content from output values. If an extractor is specified, it will be executed against each final output value. Any content which matches the extractor will be removed from the output value. If the resulting output value is empty or contains only whitespace characters, the entire output value will be discarded.

The extractor specified here MUST be match a contiguous sequence of characters within the text flow. As such, the extractor cannot use any Collation Provider Methods which combine instances geometrically.

Defines how extractor results are interpreted and transformed into the final result set.

Specifies the sort order of the result set. Can be one of the following values:

• Position - Results are ordered by their position within the content flow.
• Frequency - Results are ordered by the number of occurrences of each distinct value.
• Confidence - Results are ordered by confidence.
• Extractor - Results are ordered by the extractor which produced each match. Can be used to prioritize the results from one extractor those of over another.
• Length - Results are ordered by the length of the value.
• Value - Results are ordered by value.

Specifies whether the result set should be ordered in ascending or descending order. Can be one of the following values:

• Ascending - Results are returned in ascending order, where smaller values appear before larger values.
• Descending - Results are returned in descending order, where larger values appear before smaller values.

Defines lexicon lookup and fuzzy matching options.

Specifies options for filtering output instances.

Specifies optional processing for each output instance.

Specifies an optional post-processing operation to the applied to each output instance.

Specifies the mode to be used for deduplicating overlapping results. Can be one of the following values:

• None - No deduplication of overlapping results will be performed.
• Area - The instance occupying the largest geometric region will win. In the case of a tie, the instance with the highest confidence will win.
• Length - The instance matching the longest span will win. In the case of a tie, the instance with the highest confidence will win.
• Confidence - The instance with the highest confidence will win. In the case of a tie, the instance with the greatest length will win.
• Count - The instance matching the most characters will win. In the case of a tie, the instance with the highest confidence will win.

• In some Data Types, each extractor is attempting to match the same value using different techniques for redundancy purposes. In the event that one method fails, the other may succeed. When multiple extractors do succeed, it is can be undesirable to have multiple copies of the same result in the output. Deduplication by any method will ensure that only a single result is included in the output.
• Other Data Types may prefer a match by Extractor A over a match by Extractor B. In such cases, the extractors would each be configured to output a confidence value reflecting their relative desirability level, and results would be deduplicated by confidence.
• There are also cases where a Data Type is extracting values which are self-containing. For example, Extractor A matches "OWNERSHIP REPORT" and Extractor B matches "MINERAL OWNERSHIP REPORT". In document containing the text "MINERAL OWNERSHIP REPORT", both would successfully match. Deduplicating by length (or area) in this case would discard "OWNERSHIP REPORT" and include "MINERAL OWNERSHIP REPORT" in the output.

If True, duplicate values will be eliminated, leaving only the first instance of each distinct value. Note that deduplication is not case-sensitive.