Skip to content

Commit f93f76f

Browse files
megvanhuygenmegvanhuygen
authored
Merge pull request #102446 from Careyjmac/updates
Updates for ImageAnalysisSkill and character extraction limit
2 parents 37ca342 + d004daf commit f93f76f

9 files changed

+161
-141
lines changed

‎articles/search/cognitive-search-common-errors-warnings.md

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -299,7 +299,7 @@ The [Table Storage service](https://azure.microsoft.com/services/storage/tables)
299299
<a name="truncated-extracted-text-to-x-characters"/>
300300

301301
## Warning: Truncated extracted text to X characters
302-
Indexers limit how much text can be extracted from any one document. This limit depends on the pricing tier: 32,000 characters for Free tier, 64,000 for Basic, and 4 million for Standard, Standard S2 and Standard S3 tiers. Text that was truncated will not be indexed. To avoid this warning, try breaking apart documents with large amounts of text into multiple, smaller documents.
302+
Indexers limit how much text can be extracted from any one document. This limit depends on the pricing tier: 32,000 characters for Free tier, 64,000 for Basic, 4 million for Standard, 8 million for Standard S2, and 16 million for Standard S3. Text that was truncated will not be indexed. To avoid this warning, try breaking apart documents with large amounts of text into multiple, smaller documents.
303303

304304
For more information, see [Indexer limits](search-limits-quotas-capacity.md#indexer-limits).
305305

‎articles/search/cognitive-search-concept-image-scenarios.md

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -210,7 +210,7 @@ As a helper, if you need to transform normalized coordinates to the original coo
210210

211211
## See also
212212
+ [Create indexer (REST)](https://docs.microsoft.com/rest/api/searchservice/create-indexer)
213-
+ [Analyze image skill](cognitive-search-skill-image-analysis.md)
213+
+ [Image Analysis skill](cognitive-search-skill-image-analysis.md)
214214
+ [OCR skill](cognitive-search-skill-ocr.md)
215215
+ [Text merge skill](cognitive-search-skill-textmerger.md)
216216
+ [How to define a skillset](cognitive-search-defining-skillset.md)

‎articles/search/cognitive-search-concept-intro.md

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -104,7 +104,7 @@ Indexes are generated from an index schema that defines the fields, attributes,
104104
| Cognitive skill | An atomic transformation in an enrichment pipeline. Often, it is a component that extracts or infers structure, and therefore augments your understanding of the input data. Almost always, the output is text-based and the processing is natural language processing or image processing that extracts or generates text from image inputs. Output from a skill can be mapped to a field in an index, or used as an input for a downstream enrichment. A skill is either predefined and provided by Microsoft, or custom: created and deployed by you. | [Built-in cognitive skills](cognitive-search-predefined-skills.md) |
105105
| Data extraction | Covers a broad range of processing, but pertaining to AI enrichment, the entity recognition skill is most typically used to extract data (an entity) from a source that doesn't provide that information natively. | See [Entity Recognition Skill](cognitive-search-skill-entity-recognition.md) and [Document Extraction Skill (preview)](cognitive-search-skill-document-extraction.md)|
106106
| Image processing | Infers text from an image, such as the ability to recognize a landmark, or extracts text from an image. Common examples include OCR for lifting characters from a scanned document (JPEG) file, or recognizing a street name in a photograph containing a street sign. | See [Image Analysis Skill](cognitive-search-skill-image-analysis.md) or [OCR Skill](cognitive-search-skill-ocr.md)
107-
| Natural language processing | Text processing for insights and information about text inputs. Language detection, sentiment analysis, and key phrase extraction are skills that fall under natural language processing. | See [Key Phrase Extraction Skill](cognitive-search-skill-keyphrases.md), [Language Detection Skill](cognitive-search-skill-language-detection.md), [Text Translation Skill (preview)](cognitive-search-skill-text-translation.md), [Sentiment Analysis Skill](cognitive-search-skill-sentiment.md) |
107+
| Natural language processing | Text processing for insights and information about text inputs. Language detection, sentiment analysis, and key phrase extraction are skills that fall under natural language processing. | See [Key Phrase Extraction Skill](cognitive-search-skill-keyphrases.md), [Language Detection Skill](cognitive-search-skill-language-detection.md), [Text Translation Skill](cognitive-search-skill-text-translation.md), [Sentiment Analysis Skill](cognitive-search-skill-sentiment.md) |
108108
| Document cracking | The process of extracting or creating text content from non-text sources during indexing. Optical character recognition (OCR) is an example, but generally it refers to core indexer functionality as the indexer extracts content from application files. The data source providing source file location, and the indexer definition providing field mappings, are both key factors in document cracking. | See [Indexers overview](search-indexer-overview.md) |
109109
| Shaping | Consolidate text fragments into a larger structure, or conversely break down larger text chunks into a manageable size for further downstream processing. | See [Shaper Skill](cognitive-search-skill-shaper.md), [Text Merger Skill](cognitive-search-skill-textmerger.md), [Text Split Skill](cognitive-search-skill-textsplit.md) |
110110
| Enriched documents | A transitory internal structure, generated during processing, with final output reflected in a search index. A skillset determines which enrichments are performed. Field mappings determine which data elements are added to the index. Optionally, you can create a knowledge store to persist and explore enriched documents using tools like Storage Explorer, Power BI, or any other tool that connects to Azure Blob storage. | See [Knowledge store (preview)](knowledge-store-concept-intro.md) |

‎articles/search/cognitive-search-skill-document-extraction.md

Lines changed: 1 addition & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -13,8 +13,7 @@ ms.author: chalton
1313
# Document Extraction cognitive skill
1414

1515
> [!IMPORTANT]
16-
> This skill is currently in public preview. Preview functionality is provided without a service level agreement, and is not recommended for production workloads. For more information, see [Supplemental Terms of Use for Microsoft Azure Previews](https://azure.microsoft.com/support/legal/preview-supplemental-terms/).
17-
> The [REST API version 2019-05-06-Preview](search-api-preview.md) provides preview features. There is currently no portal or .NET SDK support.
16+
> This skill is currently in public preview. Preview functionality is provided without a service level agreement, and is not recommended for production workloads. For more information, see [Supplemental Terms of Use for Microsoft Azure Previews](https://azure.microsoft.com/support/legal/preview-supplemental-terms/). There is currently no portal or .NET SDK support.
1817
1918
The **Document Extraction** skill extracts content from a file within the enrichment pipeline. This allows you to take advantage of the document extraction step that normally happens before the skillset execution with files that may be generated by other skills.
2019

‎articles/search/cognitive-search-skill-image-analysis.md

Lines changed: 154 additions & 129 deletions
Original file line numberDiff line numberDiff line change
@@ -29,8 +29,8 @@ Parameters are case-sensitive.
2929

3030
| Parameter name | Description |
3131
|--------------------|-------------|
32-
| defaultLanguageCode | A string indicating the language to return. The service returns recognition results in a specified language. If this parameter is not specified, the default value is "en". <br/><br/>Supported languages are: <br/>*en* - English (default) <br/> *zh* - Simplified Chinese|
33-
|visualFeatures | An array of strings indicating the visual feature types to return. Valid visual feature types include: <ul><li> *categories* - categorizes image content according to a taxonomy defined in the Cognitive Services [Computer Vision documentation](https://docs.microsoft.com/azure/cognitive-services/computer-vision/category-taxonomy). </li><li> *tags* - tags the image with a detailed list of words related to the image content.</li><li>*description* - describes the image content with a complete English sentence.</li><li>*faces* - detects if faces are present. If present, generates coordinates, gender, and age.</li><li> *imageType* - detects if image is clip art or a line drawing.</li><li> *color* - determines the accent color, dominant color, and whether an image is black&white.</li><li>*adult* - detects if the image is pornographic in nature (depicts nudity or a sex act). Sexually suggestive content is also detected.</li></ul> Names of visual features are case-sensitive.|
32+
| defaultLanguageCode | A string indicating the language to return. The service returns recognition results in a specified language. If this parameter is not specified, the default value is "en". <br/><br/>Supported languages are: <br/>*en* - English (default) <br/> *es* - Spanish <br/> *ja* - Japanese <br/> *pt* - Portuguese <br/> *zh* - Simplified Chinese|
33+
| visualFeatures | An array of strings indicating the visual feature types to return. Valid visual feature types include: <ul><li>*adult* - detects if the image is pornographic in nature (depicts nudity or a sex act), or is gory (depicts extreme violence or blood). Sexually suggestive content (aka racy content) is also detected.</li><li>*brands* - detects various brands within an image, including the approximate location. The *brands* visual feature is only available in English.</li><li> *categories* - categorizes image content according to a taxonomy defined in the Cognitive Services [Computer Vision documentation](https://docs.microsoft.com/azure/cognitive-services/computer-vision/category-taxonomy). </li><li> *color* - determines the accent color, dominant color, and whether an image is black & white.</li><li>*description* - describes the image content with a complete sentence in supported languages.</li><li>*faces* - detects if faces are present. If present, generates coordinates, gender and age.</li><li> *imageType* - detects if image is clip art or a line drawing.</li><li> *objects* - detects various objects within an image, including the approximate location. The *objects* visual feature is only available in English.</li><li> *tags* - tags the image with a detailed list of words related to the image content.</li></ul> Names of visual features are case-sensitive.|
3434
| details | An array of strings indicating which domain-specific details to return. Valid visual feature types include: <ul><li>*celebrities* - identifies celebrities if detected in the image.</li><li>*landmarks* - identifies landmarks if detected in the image. </li></ul> |
3535

3636
## Skill inputs
@@ -348,138 +348,163 @@ You can define output field mappings to lower-level properties, such as just lan
348348

349349
```json
350350
{
351-
"values": [
352-
{
353-
"recordId": "1",
354-
"data": {
355-
"categories": [
356-
{
357-
"name": "abstract_",
358-
"score": 0.00390625
359-
},
360-
{
361-
"name": "people_",
362-
"score": 0.83984375,
363-
"detail": {
364-
"celebrities": [
365-
{
366-
"name": "Satya Nadella",
367-
"faceBoundingBox": [
368-
{
369-
"x": 273,
370-
"y": 309
371-
},
372-
{
373-
"x": 395,
374-
"y": 309
375-
},
376-
{
377-
"x": 395,
378-
"y": 431
379-
},
380-
{
381-
"x": 273,
382-
"y": 431
383-
}
384-
],
385-
"confidence": 0.999028444
386-
}
387-
],
388-
"landmarks": [
389-
{
390-
"name": "Forbidden City",
391-
"confidence": 0.9978346
392-
}
393-
]
394-
}
395-
}
396-
],
397-
"adult": {
398-
"isAdultContent": false,
399-
"isRacyContent": false,
400-
"adultScore": 0.0934349000453949,
401-
"racyScore": 0.068613491952419281
402-
},
403-
"tags": [
404-
{
405-
"name": "person",
406-
"confidence": 0.98979085683822632
407-
},
408-
{
409-
"name": "man",
410-
"confidence": 0.94493889808654785
411-
},
412-
{
413-
"name": "outdoor",
414-
"confidence": 0.938492476940155
415-
},
416-
{
417-
"name": "window",
418-
"confidence": 0.89513939619064331
419-
}
420-
],
421-
"description": {
422-
"tags": [
423-
"person",
424-
"man",
425-
"outdoor",
426-
"window",
427-
"glasses"
428-
],
429-
"captions": [
351+
"values": [
352+
{
353+
"recordId": "1",
354+
"data": {
355+
"categories": [
356+
{
357+
"name": "abstract_",
358+
"score": 0.00390625
359+
},
360+
{
361+
"name": "people_",
362+
"score": 0.83984375,
363+
"detail": {
364+
"celebrities": [
365+
{
366+
"name": "Satya Nadella",
367+
"faceBoundingBox": [
368+
{
369+
"x": 273,
370+
"y": 309
371+
},
372+
{
373+
"x": 395,
374+
"y": 309
375+
},
430376
{
431-
"text": "Satya Nadella sitting on a bench",
432-
"confidence": 0.48293603002174407
377+
"x": 395,
378+
"y": 431
379+
},
380+
{
381+
"x": 273,
382+
"y": 431
433383
}
434-
]
435-
},
436-
"requestId": "0dbec5ad-a3d3-4f7e-96b4-dfd57efe967d",
437-
"metadata": {
438-
"width": 1500,
439-
"height": 1000,
440-
"format": "Jpeg"
441-
},
442-
"faces": [
443-
{
444-
"age": 44,
445-
"gender": "Male",
446-
"faceBoundingBox": [
447-
{
448-
"x": 1601,
449-
"y": 395
450-
},
451-
{
452-
"x": 1653,
453-
"y": 395
454-
},
455-
{
456-
"x": 1653,
457-
"y": 447
458-
},
459-
{
460-
"x": 1601,
461-
"y": 447
462-
}
463-
]
464-
}
465-
],
466-
"color": {
467-
"dominantColorForeground": "Brown",
468-
"dominantColorBackground": "Brown",
469-
"dominantColors": [
470-
"Brown",
471-
"Black"
472384
],
473-
"accentColor": "873B59",
474-
"isBwImg": false
475-
},
476-
"imageType": {
477-
"clipArtType": 0,
478-
"lineDrawingType": 0
385+
"confidence": 0.999028444
386+
}
387+
],
388+
"landmarks": [
389+
{
390+
"name": "Forbidden City",
391+
"confidence": 0.9978346
479392
}
393+
]
480394
}
481-
}
482-
]
395+
}
396+
],
397+
"adult": {
398+
"isAdultContent": false,
399+
"isRacyContent": false,
400+
"isGoryContent": false,
401+
"adultScore": 0.0934349000453949,
402+
"racyScore": 0.068613491952419281,
403+
"goreScore": 0.08928389008070282
404+
},
405+
"tags": [
406+
{
407+
"name": "person",
408+
"confidence": 0.98979085683822632
409+
},
410+
{
411+
"name": "man",
412+
"confidence": 0.94493889808654785
413+
},
414+
{
415+
"name": "outdoor",
416+
"confidence": 0.938492476940155
417+
},
418+
{
419+
"name": "window",
420+
"confidence": 0.89513939619064331
421+
}
422+
],
423+
"description": {
424+
"tags": [
425+
"person",
426+
"man",
427+
"outdoor",
428+
"window",
429+
"glasses"
430+
],
431+
"captions": [
432+
{
433+
"text": "Satya Nadella sitting on a bench",
434+
"confidence": 0.48293603002174407
435+
}
436+
]
437+
},
438+
"requestId": "0dbec5ad-a3d3-4f7e-96b4-dfd57efe967d",
439+
"metadata": {
440+
"width": 1500,
441+
"height": 1000,
442+
"format": "Jpeg"
443+
},
444+
"faces": [
445+
{
446+
"age": 44,
447+
"gender": "Male",
448+
"faceBoundingBox": [
449+
{
450+
"x": 1601,
451+
"y": 395
452+
},
453+
{
454+
"x": 1653,
455+
"y": 395
456+
},
457+
{
458+
"x": 1653,
459+
"y": 447
460+
},
461+
{
462+
"x": 1601,
463+
"y": 447
464+
}
465+
]
466+
}
467+
],
468+
"color": {
469+
"dominantColorForeground": "Brown",
470+
"dominantColorBackground": "Brown",
471+
"dominantColors": [
472+
"Brown",
473+
"Black"
474+
],
475+
"accentColor": "873B59",
476+
"isBwImg": false
477+
},
478+
"imageType": {
479+
"clipArtType": 0,
480+
"lineDrawingType": 0
481+
},
482+
"objects": [
483+
{
484+
"rectangle": {
485+
"x": 25,
486+
"y": 43,
487+
"w": 172,
488+
"h": 140
489+
},
490+
"object": "person",
491+
"confidence": 0.931
492+
}
493+
],
494+
"brands":[
495+
{
496+
"name":"Microsoft",
497+
"rectangle":{
498+
"x":20,
499+
"y":97,
500+
"w":62,
501+
"h":52
502+
}
503+
}
504+
]
505+
}
506+
}
507+
]
483508
}
484509
```
485510

0 commit comments

Comments
 (0)