You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Commit 54aa67f
Browse filesBrowse the repository at this point in the historyBrowse files
Copy file name to clipboardExpand all lines: .speakeasy/out.openapi.yaml
+62-11Lines changed: 62 additions & 11 deletions
Original file line number
Diff line number
Diff line change
@@ -11799,6 +11799,32 @@ components:
11799
11799
required:
11800
11800
- 'data'
11801
11801
type: 'object'
11802
+
ExtractionAsset:
11803
+
additionalProperties: false
11804
+
description: 'An image or PDF sent once and referenced by its key in the `assets` map. `data_base64` is the file in standard base64 with no `data:` prefix. PDF is read by document models only.'
description: 'Set to `image` when the answer was read from an image rather than the state text.'
11841
+
description: 'Set to `image` or `document` when the answer was read from an image or a PDF rather than the state text.'
11816
11842
enum:
11817
11843
- 'state'
11818
11844
- 'image'
11845
+
- 'document'
11819
11846
type: 'string'
11820
11847
x-speakeasy-unknown-values: allow
11821
11848
span:
@@ -11842,6 +11869,7 @@ components:
11842
11869
- 'value'
11843
11870
type: 'object'
11844
11871
ExtractionDateField:
11872
+
additionalProperties: false
11845
11873
properties:
11846
11874
instructions:
11847
11875
description: 'What the field should contain.'
@@ -11872,10 +11900,11 @@ components:
11872
11900
format: 'double'
11873
11901
type: 'number'
11874
11902
source:
11875
-
description: 'Set to `image` when the answer was read from an image rather than the state text.'
11903
+
description: 'Set to `image` or `document` when the answer was read from an image or a PDF rather than the state text.'
11876
11904
enum:
11877
11905
- 'state'
11878
11906
- 'image'
11907
+
- 'document'
11879
11908
type: 'string'
11880
11909
x-speakeasy-unknown-values: allow
11881
11910
span:
@@ -11902,6 +11931,7 @@ components:
11902
11931
- 'value'
11903
11932
type: 'object'
11904
11933
ExtractionDatetimeField:
11934
+
additionalProperties: false
11905
11935
properties:
11906
11936
instructions:
11907
11937
description: 'What the field should contain.'
@@ -11933,10 +11963,11 @@ components:
11933
11963
format: 'double'
11934
11964
type: 'number'
11935
11965
source:
11936
-
description: 'Set to `image` when the answer was read from an image rather than the state text.'
11966
+
description: 'Set to `image` or `document` when the answer was read from an image or a PDF rather than the state text.'
11937
11967
enum:
11938
11968
- 'state'
11939
11969
- 'image'
11970
+
- 'document'
11940
11971
type: 'string'
11941
11972
x-speakeasy-unknown-values: allow
11942
11973
span:
@@ -11967,6 +11998,7 @@ components:
11967
11998
- 'value'
11968
11999
type: 'object'
11969
12000
ExtractionDurationField:
12001
+
additionalProperties: false
11970
12002
properties:
11971
12003
decimals:
11972
12004
description: 'Decimal places kept in `value`.'
@@ -12046,10 +12078,11 @@ components:
12046
12078
- type: 'string'
12047
12079
type: 'array'
12048
12080
source:
12049
-
description: 'Set to `image` when the answer was read from an image rather than the state text.'
12081
+
description: 'Set to `image` or `document` when the answer was read from an image or a PDF rather than the state text.'
12050
12082
enum:
12051
12083
- 'state'
12052
12084
- 'image'
12085
+
- 'document'
12053
12086
type: 'string'
12054
12087
x-speakeasy-unknown-values: allow
12055
12088
spell_confidence:
@@ -12071,6 +12104,7 @@ components:
12071
12104
- 'value'
12072
12105
type: 'object'
12073
12106
ExtractionEstimateField:
12107
+
additionalProperties: false
12074
12108
properties:
12075
12109
instructions:
12076
12110
description: 'What the field should contain.'
@@ -12165,10 +12199,11 @@ components:
12165
12199
- type: 'string'
12166
12200
type: 'array'
12167
12201
source:
12168
-
description: 'Set to `image` when the answer was read from an image rather than the state text.'
12202
+
description: 'Set to `image` or `document` when the answer was read from an image or a PDF rather than the state text.'
12169
12203
enum:
12170
12204
- 'state'
12171
12205
- 'image'
12206
+
- 'document'
12172
12207
type: 'string'
12173
12208
x-speakeasy-unknown-values: allow
12174
12209
spell_confidence:
@@ -12197,6 +12232,7 @@ components:
12197
12232
- 'value'
12198
12233
type: 'object'
12199
12234
ExtractionForecastField:
12235
+
additionalProperties: false
12200
12236
properties:
12201
12237
horizon:
12202
12238
description: 'Number of forecast steps.'
@@ -12316,6 +12352,12 @@ components:
12316
12352
confidence:
12317
12353
format: 'double'
12318
12354
type: 'number'
12355
+
confidences:
12356
+
description: 'One confidence per item of `value`, in the same order.'
12357
+
items:
12358
+
format: 'double'
12359
+
type: 'number'
12360
+
type: 'array'
12319
12361
guess:
12320
12362
description: 'The most likely value even when `value` is null because `p_null` crossed the threshold.'
12321
12363
items:
@@ -12329,10 +12371,11 @@ components:
12329
12371
format: 'double'
12330
12372
type: 'number'
12331
12373
source:
12332
-
description: 'Set to `image` when the answer was read from an image rather than the state text.'
12374
+
description: 'Set to `image` or `document` when the answer was read from an image or a PDF rather than the state text.'
12333
12375
enum:
12334
12376
- 'state'
12335
12377
- 'image'
12378
+
- 'document'
12336
12379
type: 'string'
12337
12380
x-speakeasy-unknown-values: allow
12338
12381
spans:
@@ -12366,6 +12409,7 @@ components:
12366
12409
- 'value'
12367
12410
type: 'object'
12368
12411
ExtractionListField:
12412
+
additionalProperties: false
12369
12413
properties:
12370
12414
decimals:
12371
12415
description: 'Decimal places kept in `value`.'
@@ -12410,10 +12454,11 @@ components:
12410
12454
format: 'double'
12411
12455
type: 'number'
12412
12456
source:
12413
-
description: 'Set to `image` when the answer was read from an image rather than the state text.'
12457
+
description: 'Set to `image` or `document` when the answer was read from an image or a PDF rather than the state text.'
12414
12458
enum:
12415
12459
- 'state'
12416
12460
- 'image'
12461
+
- 'document'
12417
12462
type: 'string'
12418
12463
x-speakeasy-unknown-values: allow
12419
12464
span:
@@ -12441,6 +12486,7 @@ components:
12441
12486
- 'value'
12442
12487
type: 'object'
12443
12488
ExtractionNumberField:
12489
+
additionalProperties: false
12444
12490
properties:
12445
12491
decimals:
12446
12492
description: 'Decimal places kept in `value`.'
@@ -12468,11 +12514,12 @@ components:
12468
12514
- 'type'
12469
12515
type: 'object'
12470
12516
ExtractionsRequest:
12517
+
additionalProperties: false
12471
12518
properties:
12472
12519
assets:
12473
12520
additionalProperties:
12474
-
$ref: '#/components/schemas/DecisionsImageAsset'
12475
-
description: 'Images keyed by reference string, as in TypeSafe''s API. Write a reference wherever the image belongs in `state`, `instructions`, or `criteria`; every asset must be referenced at least once. Jev reads a reference as a whole string value or inside longer text. Other Decisions models read a whole-string reference wherever they read an `image_url` part and reject a reference inside longer text.'
12521
+
$ref: '#/components/schemas/ExtractionAsset'
12522
+
description: 'Assets keyed by reference string. Write a reference wherever the asset belongs in `state`or a field `instructions`; every asset must be referenced at least once. A `state` that is exactly the sole asset''s key carries no text of its own and is read as file-only. A model that reads either text or a document per request, not both (ScaleDown Extract with a PDF), needs that form: any other text around the key is state text and is rejected with 400.'
12476
12523
type: 'object'
12477
12524
fields:
12478
12525
additionalProperties:
@@ -12613,10 +12660,11 @@ components:
12613
12660
format: 'double'
12614
12661
type: 'number'
12615
12662
source:
12616
-
description: 'Set to `image` when the answer was read from an image rather than the state text.'
12663
+
description: 'Set to `image` or `document` when the answer was read from an image or a PDF rather than the state text.'
12617
12664
enum:
12618
12665
- 'state'
12619
12666
- 'image'
12667
+
- 'document'
12620
12668
type: 'string'
12621
12669
x-speakeasy-unknown-values: allow
12622
12670
span:
@@ -12643,6 +12691,7 @@ components:
12643
12691
- 'value'
12644
12692
type: 'object'
12645
12693
ExtractionTextField:
12694
+
additionalProperties: false
12646
12695
properties:
12647
12696
instructions:
12648
12697
description: 'What the field should contain.'
@@ -12673,10 +12722,11 @@ components:
12673
12722
format: 'double'
12674
12723
type: 'number'
12675
12724
source:
12676
-
description: 'Set to `image` when the answer was read from an image rather than the state text.'
12725
+
description: 'Set to `image` or `document` when the answer was read from an image or a PDF rather than the state text.'
An image or PDF sent once and referenced by its key in the `assets` map. `data_base64` is the file in standard base64 with no `data:` prefix. PDF is read by document models only.
Copy file name to clipboardExpand all lines: docs/components/extractiondateanswer.mdx
+1-1Lines changed: 1 addition & 1 deletion
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -9,7 +9,7 @@ title: "ExtractionDateAnswer"
9
9
|`confidence`|*Optional[float]*| β | N/A |
10
10
|`guess`|*Optional[str]*| β | The most likely value even when `value` is null because `p_null` crossed the threshold. |
11
11
|`p_null`|*Optional[float]*| β | Probability that the state contains nothing for this field. |
12
-
|`source`|[Optional[components.ExtractionDateAnswerSource]](../components/extractiondateanswersource.mdx)| β | Set to `image` when the answer was read from an image rather than the state text.|
12
+
|`source`|[Optional[components.ExtractionDateAnswerSource]](../components/extractiondateanswersource.mdx)| β | Set to `image`or `document`when the answer was read from an image or a PDF rather than the state text. |
13
13
|`span`| List[*int*]| β | Unicode code point offsets [start, end) into the state text: index code points (not UTF-16 code units or UTF-8 bytes), so an emoji or accented letter counts once. |
14
14
|`spell_confidence`|*Optional[float]*| β | Confidence in the spelling of text read from an image. |
0 commit comments