Input configuration for AutoMl.ImportData action.
The format of input depends on dataset_metadata the Dataset into which
the import is happening has. As input source the
[gcs_source][google.cloud.automl.v1.InputConfig.gcs_source] is
expected, unless specified otherwise. Additionally any input .CSV file
by itself must be 100MB or smaller, unless specified otherwise. If an
"example" file (that is, image, video etc.) with identical content (even
if it had different GCS_FILE_PATH
) is mentioned multiple times, then
its label, bounding boxes etc. are appended. The same file should be
always provided with the same ML_USE
and GCS_FILE_PATH
, if it is
not, then these values are nondeterministically selected from the given
ones.
The formats are represented in EBNF with commas being literal and with non-terminal symbols defined near the end of this comment. The formats are:
See Preparing your training
data <https://cloud.google.com/vision/automl/docs/prepare>
__ for more
information.
CSV file(s) with each line in format:
::
ML_USE,GCS_FILE_PATH,LABEL,LABEL,...
ML_USE
- Identifies the data set that the current row (file) applies to. This value can be one of the following:TRAIN
- Rows in this file are used to train the model.TEST
- Rows in this file are used to test the model during training.UNASSIGNED
- Rows in this file are not categorized. They are Automatically divided into train and test data. 80% for training and 20% for testing.
GCS_FILE_PATH
- The Google Cloud Storage location of an image of up to 30MB in size. Supported extensions: .JPEG, .GIF, .PNG, .WEBP, .BMP, .TIFF, .ICO.LABEL
- A label that identifies the object in the image.
For the MULTICLASS
classification type, at most one LABEL
is
allowed per image. If an image has not yet been labeled, then it should
be mentioned just once with no LABEL
.
Some sample rows:
::
TRAIN,gs://folder/image1.jpg,daisy
TEST,gs://folder/image2.jpg,dandelion,tulip,rose
UNASSIGNED,gs://folder/image3.jpg,daisy
UNASSIGNED,gs://folder/image4.jpg
See Preparing your training
data <https://cloud.google.com/vision/automl/object-detection/docs/prepare>
__
for more information.
A CSV file(s) with each line in format:
::
ML_USE,GCS_FILE_PATH,[LABEL],(BOUNDING_BOX | ,,,,,,,)
ML_USE
- Identifies the data set that the current row (file) applies to. This value can be one of the following:TRAIN
- Rows in this file are used to train the model.TEST
- Rows in this file are used to test the model during training.UNASSIGNED
- Rows in this file are not categorized. They are Automatically divided into train and test data. 80% for training and 20% for testing.
GCS_FILE_PATH
- The Google Cloud Storage location of an image of up to 30MB in size. Supported extensions: .JPEG, .GIF, .PNG. Each image is assumed to be exhaustively labeled.LABEL
- A label that identifies the object in the image specified by theBOUNDING_BOX
.BOUNDING BOX
- The vertices of an object in the example image. The minimum allowedBOUNDING_BOX
edge length is 0.01, and no more than 500BOUNDING_BOX
instances per image are allowed (oneBOUNDING_BOX
per line). If an image has no looked for objects then it should be mentioned just once with no LABEL and the ",,,,,,," in place of theBOUNDING_BOX
.
Four sample rows:
::
TRAIN,gs://folder/image1.png,car,0.1,0.1,,,0.3,0.3,,
TRAIN,gs://folder/image1.png,bike,.7,.6,,,.8,.9,,
UNASSIGNED,gs://folder/im2.png,car,0.1,0.1,0.2,0.1,0.2,0.3,0.1,0.3
TEST,gs://folder/im3.png,,,,,,,,,
See Preparing your training
data </natural-language/automl/entity-analysis/docs/prepare>
__ for more
information.
One or more CSV file(s) with each line in the following format:
::
ML_USE,GCS_FILE_PATH
ML_USE
- Identifies the data set that the current row (file) applies to. This value can be one of the following:TRAIN
- Rows in this file are used to train the model.TEST
- Rows in this file are used to test the model during training.UNASSIGNED
- Rows in this file are not categorized. They are Automatically divided into train and test data. 80% for training and 20% for testing..
GCS_FILE_PATH
- a Identifies JSON Lines (.JSONL) file stored in Google Cloud Storage that contains in-line text in-line as documents for model training.
After the training data set has been determined from the TRAIN
and
UNASSIGNED
CSV files, the training data is divided into train and
validation data sets. 70% for training and 30% for validation.
For example:
::
TRAIN,gs://folder/file1.jsonl
VALIDATE,gs://folder/file2.jsonl
TEST,gs://folder/file3.jsonl
In-line JSONL files
In-line .JSONL files contain, per line, a JSON document that wraps a
[text_snippet
][google.cloud.automl.v1.TextSnippet] field followed by
one or more [annotations
][google.cloud.automl.v1.AnnotationPayload]
fields, which have display_name
and text_extraction
fields to
describe the entity from the text snippet. Multiple JSON documents can
be separated using line breaks (\n
).
The supplied text must be annotated exhaustively. For example, if you include the text "horse", but do not label it as "animal", then "horse" is assumed to not be an "animal".
Any given text snippet content must have 30,000 characters or less, and also be UTF-8 NFC encoded. ASCII is accepted as it is UTF-8 NFC encoded.
For example:
::
{
"text_snippet": {
"content": "dog car cat"
},
"annotations": [
{
"display_name": "animal",
"text_extraction": {
"text_segment": {"start_offset": 0, "end_offset": 2}
}
},
{
"display_name": "vehicle",
"text_extraction": {
"text_segment": {"start_offset": 4, "end_offset": 6}
}
},
{
"display_name": "animal",
"text_extraction": {
"text_segment": {"start_offset": 8, "end_offset": 10}
}
}
]
}\n
{
"text_snippet": {
"content": "This dog is good."
},
"annotations": [
{
"display_name": "animal",
"text_extraction": {
"text_segment": {"start_offset": 5, "end_offset": 7}
}
}
]
}
JSONL files that reference documents
.JSONL files contain, per line, a JSON document that wraps a
input_config
that contains the path to a source PDF document.
Multiple JSON documents can be separated using line breaks
(\n
).
For example:
::
{
"document": {
"input_config": {
"gcs_source": { "input_uris": [ "gs://folder/document1.pdf" ]
}
}
}
}\n
{
"document": {
"input_config": {
"gcs_source": { "input_uris": [ "gs://folder/document2.pdf" ]
}
}
}
}
In-line JSONL files with PDF layout information
Note: You can only annotate PDF files using the UI. The format
described below applies to annotated PDF files exported using the UI or
exportData
.
In-line .JSONL files for PDF documents contain, per line, a JSON
document that wraps a document
field that provides the textual
content of the PDF document and the layout information.
For example:
::
{
"document": {
"document_text": {
"content": "dog car cat"
}
"layout": [
{
"text_segment": {
"start_offset": 0,
"end_offset": 11,
},
"page_number": 1,
"bounding_poly": {
"normalized_vertices": [
{"x": 0.1, "y": 0.1},
{"x": 0.1, "y": 0.3},
{"x": 0.3, "y": 0.3},
{"x": 0.3, "y": 0.1},
],
},
"text_segment_type": TOKEN,
}
],
"document_dimensions": {
"width": 8.27,
"height": 11.69,
"unit": INCH,
}
"page_count": 3,
},
"annotations": [
{
"display_name": "animal",
"text_extraction": {
"text_segment": {"start_offset": 0, "end_offset": 3}
}
},
{
"display_name": "vehicle",
"text_extraction": {
"text_segment": {"start_offset": 4, "end_offset": 7}
}
},
{
"display_name": "animal",
"text_extraction": {
"text_segment": {"start_offset": 8, "end_offset": 11}
}
},
],
See Preparing your training
data <https://cloud.google.com/natural-language/automl/docs/prepare>
__
for more information.
One or more CSV file(s) with each line in the following format:
::
ML_USE,(TEXT_SNIPPET | GCS_FILE_PATH),LABEL,LABEL,...
ML_USE
- Identifies the data set that the current row (file) applies to. This value can be one of the following:TRAIN
- Rows in this file are used to train the model.TEST
- Rows in this file are used to test the model during training.UNASSIGNED
- Rows in this file are not categorized. They are Automatically divided into train and test data. 80% for training and 20% for testing.
TEXT_SNIPPET
andGCS_FILE_PATH
are distinguished by a pattern. If the column content is a valid Google Cloud Storage file path, that is, prefixed by "gs://", it is treated as aGCS_FILE_PATH
. Otherwise, if the content is enclosed in double quotes (""), it is treated as aTEXT_SNIPPET
. ForGCS_FILE_PATH
, the path must lead to a file with supported extension and UTF-8 encoding, for example, "gs://folder/content.txt" AutoML imports the file content as a text snippet. ForTEXT_SNIPPET
, AutoML imports the column content excluding quotes. In both cases, size of the content must be 10MB or less in size. For zip files, the size of each file inside the zip must be 10MB or less in size.For the
MULTICLASS
classification type, at most oneLABEL
is allowed. TheML_USE
andLABEL
columns are optional. Supported file extensions: .TXT, .PDF, .ZIP
A maximum of 100 unique labels are allowed per CSV row.
Sample rows:
::
TRAIN,"They have bad food and very rude",RudeService,BadFood
gs://folder/content.txt,SlowService
TEST,gs://folder/document.pdf
VALIDATE,gs://folder/text_files.zip,BadFood
See Preparing your training
data <https://cloud.google.com/natural-language/automl/docs/prepare>
__
for more information.
CSV file(s) with each line in format:
::
ML_USE,(TEXT_SNIPPET | GCS_FILE_PATH),SENTIMENT
ML_USE
- Identifies the data set that the current row (file) applies to. This value can be one of the following:TRAIN
- Rows in this file are used to train the model.TEST
- Rows in this file are used to test the model during training.UNASSIGNED
- Rows in this file are not categorized. They are Automatically divided into train and test data. 80% for training and 20% for testing.
TEXT_SNIPPET
andGCS_FILE_PATH
are distinguished by a pattern. If the column content is a valid Google Cloud Storage file path, that is, prefixed by "gs://", it is treated as aGCS_FILE_PATH
. Otherwise, if the content is enclosed in double quotes (""), it is treated as aTEXT_SNIPPET
. ForGCS_FILE_PATH
, the path must lead to a file with supported extension and UTF-8 encoding, for example, "gs://folder/content.txt" AutoML imports the file content as a text snippet. ForTEXT_SNIPPET
, AutoML imports the column content excluding quotes. In both cases, size of the content must be 128kB or less in size. For zip files, the size of each file inside the zip must be 128kB or less in size.The
ML_USE
andSENTIMENT
columns are optional. Supported file extensions: .TXT, .PDF, .ZIPSENTIMENT
- An integer between 0 and Dataset.text_sentiment_dataset_metadata.sentiment_max (inclusive). Describes the ordinal of the sentiment - higher value means a more positive sentiment. All the values are completely relative, i.e. neither 0 needs to mean a negative or neutral sentiment nor sentiment_max needs to mean a positive one - it is just required that 0 is the least positive sentiment in the data, and sentiment_max is the most positive one. The SENTIMENT shouldn't be confused with "score" or "magnitude" from the previous Natural Language Sentiment Analysis API. All SENTIMENT values between 0 and sentiment_max must be represented in the imported data. On prediction the same 0 to sentiment_max range will be used. The difference between neighboring sentiment values needs not to be uniform, e.g. 1 and 2 may be similar whereas the difference between 2 and 3 may be large.
Sample rows:
::
TRAIN,"@freewrytin this is way too good for your product",2
gs://folder/content.txt,3
TEST,gs://folder/document.pdf
VALIDATE,gs://folder/text_files.zip,2
Input field definitions:
ML_USE
("TRAIN" | "VALIDATE" | "TEST" | "UNASSIGNED") Describes how the
given example (file) should be used for model training. "UNASSIGNED"
can be used when user has no preference.
GCS_FILE_PATH
The path to a file on Google Cloud Storage. For example,
"gs://folder/image1.png".
LABEL
A display name of an object on an image, video etc., e.g. "dog".
Must be up to 32 characters long and can consist only of ASCII Latin
letters A-Z and a-z, underscores(_), and ASCII digits 0-9. For each
label an AnnotationSpec is created which display_name becomes the
label; AnnotationSpecs are given back in predictions.
BOUNDING_BOX
(VERTEX,VERTEX,VERTEX,VERTEX
| VERTEX,,,VERTEX,,
) A
rectangle parallel to the frame of the example (image, video). If 4
vertices are given they are connected by edges in the order
provided, if 2 are given they are recognized as diagonally opposite
vertices of the rectangle.
VERTEX
(COORDINATE,COORDINATE
) First coordinate is horizontal (x), the
second is vertical (y).
COORDINATE
A float in 0 to 1 range, relative to total length of image or video
in given dimension. For fractions the leading non-decimal 0 can be
omitted (i.e. 0.3 = .3). Point 0,0 is in top left.
TEXT_SNIPPET
The content of a text snippet, UTF-8 encoded, enclosed within double
quotes ("").
DOCUMENT
A field that provides the textual content with document and the
layout information.
Errors:
If any of the provided CSV files can't be parsed or if more than certain percent of CSV rows cannot be processed then the operation fails and nothing is imported. Regardless of overall success or failure the per-row failures, up to a certain count cap, is listed in Operation.metadata.partial_failures.
The Google Cloud Storage location for the input content. For
AutoMl.ImportData,
gcs_source
points to a CSV file with a structure described
in InputConfig.
Classes
ParamsEntry
API documentation for automl_v1.types.InputConfig.ParamsEntry
class.