com.clarifai.grpc.api.AndOrBuilder Maven / Gradle / Ivy
The newest version!
// Generated by the protocol buffer compiler. DO NOT EDIT!
// source: proto/clarifai/api/resources.proto
package com.clarifai.grpc.api;
public interface AndOrBuilder extends
// @@protoc_insertion_point(interface_extends:clarifai.api.And)
com.google.protobuf.MessageOrBuilder {
/**
*
* FILTER by input.data... information.
* This can include human provided concepts, geo location info, metadata, etc.
* This is effectively searching over only the trusted annotation attached to an input in your
* app. To search by more specific annotation fields use the Annotation object here.
* ########## Supported fields ##########
* - data.concepts[].id
* - data.concepts[].name
* - data.concepts[].value
* - data.geo.geo_box[].geo_point.latitude
* - data.geo.geo_box[].geo_point.longitude
* - data.geo.geo_limit.type
* - data.geo.geo_limit.value
* - data.geo.geo_point.latitude
* - data.geo.geo_point.longitude
* - data.image.url
* - data.metadata.fields - filter by metadata. metadata key&value fields are OR-ed.
* - dataset_ids[] - filter by dataset IDs
* - id - filter by input ID
* - status.code - filter by input status
*
*
* .clarifai.api.Input input = 1;
* @return Whether the input field is set.
*/
boolean hasInput();
/**
*
* FILTER by input.data... information.
* This can include human provided concepts, geo location info, metadata, etc.
* This is effectively searching over only the trusted annotation attached to an input in your
* app. To search by more specific annotation fields use the Annotation object here.
* ########## Supported fields ##########
* - data.concepts[].id
* - data.concepts[].name
* - data.concepts[].value
* - data.geo.geo_box[].geo_point.latitude
* - data.geo.geo_box[].geo_point.longitude
* - data.geo.geo_limit.type
* - data.geo.geo_limit.value
* - data.geo.geo_point.latitude
* - data.geo.geo_point.longitude
* - data.image.url
* - data.metadata.fields - filter by metadata. metadata key&value fields are OR-ed.
* - dataset_ids[] - filter by dataset IDs
* - id - filter by input ID
* - status.code - filter by input status
*
*
* .clarifai.api.Input input = 1;
* @return The input.
*/
com.clarifai.grpc.api.Input getInput();
/**
*
* FILTER by input.data... information.
* This can include human provided concepts, geo location info, metadata, etc.
* This is effectively searching over only the trusted annotation attached to an input in your
* app. To search by more specific annotation fields use the Annotation object here.
* ########## Supported fields ##########
* - data.concepts[].id
* - data.concepts[].name
* - data.concepts[].value
* - data.geo.geo_box[].geo_point.latitude
* - data.geo.geo_box[].geo_point.longitude
* - data.geo.geo_limit.type
* - data.geo.geo_limit.value
* - data.geo.geo_point.latitude
* - data.geo.geo_point.longitude
* - data.image.url
* - data.metadata.fields - filter by metadata. metadata key&value fields are OR-ed.
* - dataset_ids[] - filter by dataset IDs
* - id - filter by input ID
* - status.code - filter by input status
*
*
* .clarifai.api.Input input = 1;
*/
com.clarifai.grpc.api.InputOrBuilder getInputOrBuilder();
/**
*
* RANK based predicted outputs from models such as custom trained models, pre-trained models,
* etc. This is also where you enter the image url for a visual search because what we're asking
* the system to do is find output embedding most visually similar to the provided input (that
* input being in And.output.input.data.image.url for example). This will return the Hits
* sorted by visual similarity (1.0 being very similar or exact match and 0.0 being very
* dissimlar). For a search by Output concept, this means we're asking the system to rank
* the Hits by confidence of our model's predicted Outputs. So for example if the model
* predicts an image is 0.95 likely there is a "dog" present, that should related directly
* to the score returned if you search for Output concept "dog" in your query. This provides
* a natural ranking to search results based on confidence of predictions from the models and
* is used when ANDing multiple of these types of RANK by Output queries together as well.
* ########## Supported fields ##########
* - data.clusters[].id
* - data.concepts[].id
* - data.concepts[].name
* - data.concepts[].value
* - input.data.image.base64[]
* - input.data.image.url
* - input.id
*
*
* .clarifai.api.Output output = 2;
* @return Whether the output field is set.
*/
boolean hasOutput();
/**
*
* RANK based predicted outputs from models such as custom trained models, pre-trained models,
* etc. This is also where you enter the image url for a visual search because what we're asking
* the system to do is find output embedding most visually similar to the provided input (that
* input being in And.output.input.data.image.url for example). This will return the Hits
* sorted by visual similarity (1.0 being very similar or exact match and 0.0 being very
* dissimlar). For a search by Output concept, this means we're asking the system to rank
* the Hits by confidence of our model's predicted Outputs. So for example if the model
* predicts an image is 0.95 likely there is a "dog" present, that should related directly
* to the score returned if you search for Output concept "dog" in your query. This provides
* a natural ranking to search results based on confidence of predictions from the models and
* is used when ANDing multiple of these types of RANK by Output queries together as well.
* ########## Supported fields ##########
* - data.clusters[].id
* - data.concepts[].id
* - data.concepts[].name
* - data.concepts[].value
* - input.data.image.base64[]
* - input.data.image.url
* - input.id
*
*
* .clarifai.api.Output output = 2;
* @return The output.
*/
com.clarifai.grpc.api.Output getOutput();
/**
*
* RANK based predicted outputs from models such as custom trained models, pre-trained models,
* etc. This is also where you enter the image url for a visual search because what we're asking
* the system to do is find output embedding most visually similar to the provided input (that
* input being in And.output.input.data.image.url for example). This will return the Hits
* sorted by visual similarity (1.0 being very similar or exact match and 0.0 being very
* dissimlar). For a search by Output concept, this means we're asking the system to rank
* the Hits by confidence of our model's predicted Outputs. So for example if the model
* predicts an image is 0.95 likely there is a "dog" present, that should related directly
* to the score returned if you search for Output concept "dog" in your query. This provides
* a natural ranking to search results based on confidence of predictions from the models and
* is used when ANDing multiple of these types of RANK by Output queries together as well.
* ########## Supported fields ##########
* - data.clusters[].id
* - data.concepts[].id
* - data.concepts[].name
* - data.concepts[].value
* - input.data.image.base64[]
* - input.data.image.url
* - input.id
*
*
* .clarifai.api.Output output = 2;
*/
com.clarifai.grpc.api.OutputOrBuilder getOutputOrBuilder();
/**
*
* If True then this will flip the meaning of this part of the
* query. This allow for queries such as dog AND ! metadata=={"blah":"value"}
*
*
* bool negate = 3;
* @return The negate.
*/
boolean getNegate();
/**
*
* FILTER by annotation information. This is more flexible than just filtering by
* Input information because in the general case each input can have several annotations.
* Some example use cases for filtering by annotations:
* 1) find all the inputs annotated "dog" by worker_id = "XYZ"
* 2) find all the annotations associated with embed_model_version_id = "123"
* 3) find all the annotations that are trusted, etc.
* Since all the annotations under the hood are joined to the embedding model's annotation
* using worker_id's of other models like cluster models or concept models should be
* combinable with queries like visual search (a query with Output filled in).
* ########## Supported fields ##########
* - annotation_info.fields - filter by annotation info
* - data.concepts[].id
* - data.concepts[].name
* - data.concepts[].value
* - data.geo.geo_box[].geo_point.latitude
* - data.geo.geo_box[].geo_point.longitude
* - data.geo.geo_limit.type
* - data.geo.geo_limit.value
* - data.geo.geo_point.latitude
* - data.geo.geo_point.longitude
* - data.image.url
* - data.metadata.fields - filter by metadata. metadata key&value fields are OR-ed.
* - input_id
* - input_level
* - model_version_id
* - status.code
* - task_id
* - trusted
* - user_id
*
*
* .clarifai.api.Annotation annotation = 4;
* @return Whether the annotation field is set.
*/
boolean hasAnnotation();
/**
*
* FILTER by annotation information. This is more flexible than just filtering by
* Input information because in the general case each input can have several annotations.
* Some example use cases for filtering by annotations:
* 1) find all the inputs annotated "dog" by worker_id = "XYZ"
* 2) find all the annotations associated with embed_model_version_id = "123"
* 3) find all the annotations that are trusted, etc.
* Since all the annotations under the hood are joined to the embedding model's annotation
* using worker_id's of other models like cluster models or concept models should be
* combinable with queries like visual search (a query with Output filled in).
* ########## Supported fields ##########
* - annotation_info.fields - filter by annotation info
* - data.concepts[].id
* - data.concepts[].name
* - data.concepts[].value
* - data.geo.geo_box[].geo_point.latitude
* - data.geo.geo_box[].geo_point.longitude
* - data.geo.geo_limit.type
* - data.geo.geo_limit.value
* - data.geo.geo_point.latitude
* - data.geo.geo_point.longitude
* - data.image.url
* - data.metadata.fields - filter by metadata. metadata key&value fields are OR-ed.
* - input_id
* - input_level
* - model_version_id
* - status.code
* - task_id
* - trusted
* - user_id
*
*
* .clarifai.api.Annotation annotation = 4;
* @return The annotation.
*/
com.clarifai.grpc.api.Annotation getAnnotation();
/**
*
* FILTER by annotation information. This is more flexible than just filtering by
* Input information because in the general case each input can have several annotations.
* Some example use cases for filtering by annotations:
* 1) find all the inputs annotated "dog" by worker_id = "XYZ"
* 2) find all the annotations associated with embed_model_version_id = "123"
* 3) find all the annotations that are trusted, etc.
* Since all the annotations under the hood are joined to the embedding model's annotation
* using worker_id's of other models like cluster models or concept models should be
* combinable with queries like visual search (a query with Output filled in).
* ########## Supported fields ##########
* - annotation_info.fields - filter by annotation info
* - data.concepts[].id
* - data.concepts[].name
* - data.concepts[].value
* - data.geo.geo_box[].geo_point.latitude
* - data.geo.geo_box[].geo_point.longitude
* - data.geo.geo_limit.type
* - data.geo.geo_limit.value
* - data.geo.geo_point.latitude
* - data.geo.geo_point.longitude
* - data.image.url
* - data.metadata.fields - filter by metadata. metadata key&value fields are OR-ed.
* - input_id
* - input_level
* - model_version_id
* - status.code
* - task_id
* - trusted
* - user_id
*
*
* .clarifai.api.Annotation annotation = 4;
*/
com.clarifai.grpc.api.AnnotationOrBuilder getAnnotationOrBuilder();
}