SymetryML Documentation

Data Source API

Using a Data Source

In SymetryML, a data source is an abstraction of a CSV file that resides somewhere and that can be used by:

  • SymetryML projects to learn new data
  • Models to make predictions and assessments
  • Encoder to update their internal encoding table

SymetryML supports various types of data sources:

  • Secure File Transfer Protocol (SFTP)
  • HTTP/HTTPS URL
  • Amazon Simple Storage Service (S3)
  • Microsoft Azure Blob Storage
  • Google Cloud Storage
  • Oracle OCI Object Storage
  • Amazon RedShift
  • Snowflake
  • Spark Processing: Amazon S3, Google cloud storage, Oracle OCI object storage and Microsoft Azure blob storage data sources can be processed in parallel leveraging a Spark Cluster.
  • SymetryML data source plugins
  • JDBC
  • Local Data Source, that is allows to browse the local file system of the jetty web server with same privileges as the user running the Jetty web server.

Streaming Data Sources: For real-time streaming data sources such as Kafka and NATS, please refer to the Stream Data Source API documentation.

To use a data source, create a JSON data structure described in DSInfo that contains the fields in table below.

Required DSInfo Fields

FieldDescription
type

Type of Data Source

- Secure FTP (SFTP) data source = sftp

- HTTP/HTTPS data source = http

- Amazon S3 = s3

- Oracle OCI Object Storage with S3 Compatibility = s3oci

- Google Clound Storage = gcs

- Amazon Redshift = redshift

- Spark Data Source = please see the Spark Data Source type section for the matrix of all possible data source names involving Spark Processing.

- Data Source Plug ins. - jdbc

- Snowflake = snow

- Local file = localfile

- Amazon Elastic Map Reduce = emr

- Microsoft Azure Blob Storage = abs

- Streaming Data Sources (Kafka, NATS) = kafka or nats - see Stream Data Source API

nameName of the data source.
infoHash Map Containing Additional Information Based on Data Source. Please consult the next section for details about this field.

Additional Information Stored in Data Source

The info field of a data source contains specific information based on the type of data source. The following tables describes what they are for each different type of data source

HTTP/HTTPS Data Source

fieldDescription
pathhttp:// or https:// URL.

Secure FTP (SFTP)

fieldDescription
pathpath to the file on the server.
sftpuseruser name used to connect to the SFTP server.
sftppasswduser password used to connect to the SFTP server.
sftphosthost to which you want to connect.

Amazon S3

fieldDescription
pathpath to file on the server, excluding the Amazon S3 bucket
s3accessKeyAmazon S3 access key to use to connect to S3.
s3secretKeyAmazon S3 secret key to use to connect to S3.
s3bucketAmazon S3 bucket to use.

Oracle OCI Object Storage with S3 Compatibility

Please consult https://docs.oracle.com/en-us/iaas/Content/Object/Tasks/s3compatibleapi.htm for details on how to configure your Oracle OCI account so that the Amazon S3 Compatibility API can be used.

fieldDescription
pathpath to file on the server, excluding the Oracle OCI Object Storage bucket
s3accessKeyOracle Oracle OCI Object Storage access key to use to connect to Oracle OCI Object Storage.
s3secretKeyOracle Oracle OCI Object Storage secret key to use to connect to Oracle OCI Object Storage.
s3bucketOracle Oracle OCI Object Storage bucket to use.
ocinamespaceOracle Oracle OCI Object Storage namespace
ociregionOracle Oracle OCI Object Storage region

Google Cloud Storage

fieldDescription
pathPath to file
gcsaccessKeyGCS HMAC Access Key
gcssecretKeyGCS HMAC Secret Key
gcsbucketGCS Bucket
gcsprojectGCS Project
gcsmarkerOptional marker parameter indicating where in the GCS bucket to begin listing. The list will only include keys that occur lexicographically after the marker.
gcsdelimiterGCS File/Folder delimiter. / is used by default

Microsoft Azure Blob Storage

fieldDescription
pathPath to file
azure.credentials.connection.stringConnection string that specifies credentials to authorize access to Azure Blob Storage. Use one of this, account key or SAS token.
azure.account.nameName of the Azure account to use
azure.credentials.sharedkey.account.keyAccount key that specifies credentials to authorize access to Azure Blob Storage. Use one of this, connection string or SAS token.
azure.credentials.sharedkey.sas.tokenSAS token (account or service) that specifies credentials to authorize access to Azure Blob Storage. Use one of this, connection string or account key.
azure.blob.container.nameName of the Azure Blob Storage container that contains the blob
azure.blob.inputstream.chunk.size.max.bytesMaximum size in bytes of each chunk of data when reading the blob contents chunk by chunk. Default: 4194304
azure.blob.path.delimiterString that separates elements of the path to the blob file. Default: /
azure.blob.list.marker

Marker that specifies the beginning of the next page of a list of
Azure Blob Storage items to fetch from Azure. This property is used internally.

Amazon Redshift

fieldDescription
pathname of the table to use.
rsuserRedshift database user
rspasswdRedshift user password
rsurlRedshift connection url

Spark Map Reduce

fieldDescription
sparkmasteraddress of the spark’s cluster master
spark.job.process.jvm.heap.size.minMininum JVM size used for the spark Driver process launched by the Jetty Rest Server. Default : 1024m
spark.job.process.jvm.heap.size.maxMaximum JVM size used for the spark Driver process launched by the Jetty Rest Server. Default: 2048m
Any Spark parameters be used also like: spark.executor.memory or spark.executor.coresTo pass such parameters, prefix them with ‘sml.sparkenv.’ as in the following examples: - sml.sparkenv.spark.executor.cores - sml.sparkenv.spark.cores.max
spark.automl.sample.random.seedIf AutoML is used, one can set the randomizer seed that will be used to select a random sample of tuple from the data source to be used to bootstrap the AutoML environement.

Spark Map Reduce Data Source Type

SymetryML 6.3.0 and later: Due to changes in the underlying JDK and Jetty server, only Spark 4.1.0 is supported.

Data SourceSpark 4.1.0
Oracle OCIs3sparkocis3_map_reduce_4_1_0
Amazon S3sparks3_map_reduce_4_1_0
Google Cloud Storagesparkgcs_map_reduce_4_1_0
Microsoft Azure Blobsparkabs_map_reduce_4_1_0

JDBC

fieldDescription
driver
host
port
database
user
password

Snowflake

SymetryML supports connecting to Snowflake data warehouses using Snowflake's Snowpark Container Services (CP3O) for distributed processing.

fieldRequiredDescription
snowflake.accountRequiredSnowflake account identifier (e.g., XYZZZZZ-ZBC00000)
snowflake.account.roleRequiredRole to connect to the Snowflake account with (e.g., ACCOUNTADMIN)
snowflake.userRequiredUser to connect to the Snowflake account with (e.g., zzzuser)
snowflake.auth.private.key.rawRequiredString content of the private key to authorize the connection to the account
snowflake.data.databaseRequiredName of the database containing the data
snowflake.data.schemaRequiredName of the schema containing the data
snowflake.data.tableRequiredName of the table containing the data
snowflake.cp3o.compute.pool.instance.familyRequiredCompute pool instance family to use (e.g., CPU_X64_S)
snowflake.cp3o.compute.pool.max.nodesRequiredMaximum number of nodes in the compute pool (e.g., 3)
snowflake.cp3o.compute.pool.min.nodesRequiredMinimum number of nodes in the compute pool (e.g., 1)
snowflake.cp3o.worker.totalRequiredNumber of workers that share the update task (e.g., 3)
snowflake.cp3o.data.table.order.by.clauseRequiredSQL clause used to order the rows of the table so that the workers can partition it properly (e.g., ROW_INDEX ASC)
snowflake.job.warehouseOptionalSnowflake warehouse to use for queries. Default: WH_NB
snowflake.job.warehouse.sizeOptionalSize of the Snowflake warehouse to use for queries. Default: XSMALL
snowflake.job.warehouse.auto.suspendOptionalInactivity timeout in seconds of the Snowflake warehouse to use for queries. Default: 120
snowflake.job.roleOptionalSnowflake role to use in processing the job. Default: C3PO_APP_ROLE
snowflake.list.rolesOptionalBoolean flag to show a list of Snowflake roles while browsing. Default: false
snowflake.cp3o.data.frame.max.sizeOptionalMaximum size of a Data Frame to use while updating SML projects in processing the job. Default: 16384

Amazon EMR

fieldRequiredDescription
chunksizeOptionalSymetryML process the data chunk by chunk. This parameters specifies the chunk size. Default: 5000
emr.client.aws.regionOptionalAWS region of the EMR cluster. Default: us-east-1
emr.cluster.ec2.key.nameRequiredEC2 key pair name for the cluster.
emr.cluster.ec2.subnet.idOptionalEC2 subnet id for the cluster. Default: null
emr.cluster.instance.countRequired# of EC2 instances in the EMR cluster.
emr.cluster.instance.master.typeRequiredInstance type of the master EC2 instance.
emr.cluster.instance.slave.typeRequiredInstance type of the slave EC2 instances.
emr.cluster.log.storage.enableOptionalBoolean enabling for storing the EMR logs. Default: false
emr.cluster.log.storage.uriOptionalURI of the EMR logs. Default: null
emr.job.flow.roleOptionalEMR role for EC2 that is used by EC2 instances within the cluster. Default: AWS EMR_EC2_DefaultRole
emr.s3.job.bucket.nameRequiredS3 bucket that stores the files needed for the Spark cluster job; it can include the directory that stores the EMR logs.
emr.service.roleOptionalAmazon EMR role, which defines the allowable actions for Amazon EMR. Default: AWS EMR_DefaultRole
pathRequiredPath of data source to process, can be a folder. This is the 'data path' without the 'bucket part'
s3accessKeyRequiredAWS access key
s3bucketRequiredAWS S3 bucket where data resides
s3markerRequiredOptional marker parameter indicating where in the S3 bucket to begin listing. The list will only include keys that occur lexicographically after the marker.
s3secretKeyRequiredAWS secret key
sml.sparkenv.*RequiredAllows to specify any Apache Spark environment configuration like: spark.cores.max e.g. use:sml.sparkenv.spark.cores.max or spark.executor.memory e.g. use:sml.sparkenv.spark.executor.memory
sparksymprojectRequiredName of the project.

Additional CSV Options

You can specify additional parameters that describe the ‘type of csv files’. You can also add the following parameters to a data source to change how SymetryML parses your data:

KeyDescription
csv_entry_separatorSpecifies which character to use as the delimiter for each record for a given tuple.
csv_quote_characterSpecifies the quote character.
csv_strict_quotesSetting this option to true discards characters outside the quotes. If there are no quotes between delimiters, an empty string is generated.
csv_header_missingSpecifies that this data source does not have any header. SymetryML can then generate a header automatically.

Additional Information on Spark S3 Data Source

SymetryML can leverage a spark cluster to speed up processing of large amounts of data significantly. Currently, your data must reside on Amazon S3. Depending on the size of your data, it may take more or less time for the job to start, as the Spark Cluster must compute the partitions of your data before starting the job. Consequently, if your data is very large, this may take a few minutes.

Please be sure to consult the Spark Installation Guide for additional Spark information.

Best practices for Spark S3 Data Source:

  • Performance may vary depending on Amazon resource utilization when you run your job.
  • Be sure all executor nodes in your cluster reside in the same Amazon EC2 placement group.

About Data Source Plugins (DSPlugins)

The SymetryML data source API allows you to create a new data source by the form of Java library (jar) that can be added the server. Instead of transforming data into CSV files, for example, you can write a DS plugin that reads the data natively.

Data Source Encryption

Data sources might contain sensitive information that should never be passed in the clear. To avoid having to use HTTPS for these services, the SymetryML REST API forces you to pass such information in encrypted form. This can be done easily, as each SymetryML secret key is also a 128-bit Advanced Encryption Standard (AES) secret key.

  1. Create your DSInfo data structure and enter the appropriate information so that the SymetryML server can access it.
  2. Extract the JSON string from that data structure.
  3. Encrypt the JSON string representation using your SymetryML secret key:
    • Initialization vector in Base 64: LzM5QUtXZXWHm7HJ4wAePg==
    • Block cipher algorithm: AES/CBC/PKCS5Padding
  4. Send the encrypted string as part of the body to any REST service that requests a DSInfo as the body.
  5. The server decrypts the string using the client secret key and reconstructs the DSInfo.

Data Source Create

This API function creates a new data source.

URL

POST /symetry/rest/{cid}/dss/create [Body=DSInfo (encrypted)]

HTTP Responses

HTTP Status CodeHTTP Status MessageDescription
202CREATEDSuccess.
409CONFLICTA data source with the specified name already exists.

HTTP Response Entity

None.

Sample Request/Response

Please note that in the following example the BODY is encrypted. Refer to section Data Source Encryption for encryption details.

Request:
POST url="http://charm:8080/symetry/rest/c1/dss/create"

Body:
emUmJ7LXKaICPww/dKRoMR/Go6+B3ATTn+GwgE1/vcK9pN/mLoqGyKiMtGiTYMct4Gth1ElniKLXtmXfHfs7Rfn+QhJHZ+s00w2PBdbvYZIoF3My04H5XCboY21Fh4SkBhsxo+DhYuardN7R+uGphN/DSbiHRLIXe51HijSpfuq0fJuQYW0ccN4FM/B74LJccuDnbb+IouS9u/9rybKu/wsVbDVRTK/zZpWXyn4qtds=

Response:
{"statusCode":"CREATED","statusString":"DS created with name [ds1] for customer [c1]","values":{}}

Data Source Update

This API function update an existing data source.

URL

PUT /symetry/rest/{cid}/dss/{dsname} [Body=DSInfo (encrypted)]

HTTP Responses

HTTP Status CodeHTTP Status MessageDescription
200OKSuccess.
404NOT FOUNDIf data source with the specified name does not exist.

HTTP Response Entity

None.

Sample Request/Response

Please note that in the following example the BODY is encrypted. Refer to section Data Source Encryption for encryption details.

Request:
PUT url="http://charm:8080/symetry/rest/c1/dss/aDataSourceName"

Body:
emUmJ7LXKaICPww/dKRoMR/Go6+B3ATTn+GwgE1/vcK9pN/mLoqGyKiMtGiTYMct4Gth1ElniKLXtmXfHfs7Rfn+QhJHZ+s00w2PBdbvYZIoF3My04H5XCboY21Fh4SkBhsxo+DhYuardN7R+uGphN/DSbiHRLIXe51HijSpfuq0fJuQYW0ccN4FM/B74LJccuDnbb+IouS9u/9rybKu/wsVbDVRTK/zZpWXyn4qtds=

Response:
{"statusCode":"OK","statusString":"DS updated with name [ds1] for customer [c1]","values":{}}

List Customer Data Sources

This API function returns all the data sources that belong to a user.

URL

GET /symetry/rest/{cid}/dss/

HTTP Responses

HTTP Status CodeHTTP Status MessageDescription
200OKSuccess.

HTTP Response Entity

HTTP Response EntityExample
StringList{"statusCode":"OK","statusString":"OK","values":{"stringList":{"values":["Iris_SymetryML.csv-predict.csv:s3","h1:http","BigData11g_Test.csv:http","Iris_SymetryML.csv:s3","Smaato_Bids_20130812_CTR.csv:s3"]}}}

Sample Request/Response

Request:
GET url="http://charm:8080/symetry/rest/c1/dss"

Response:=
{"statusCode":"OK","statusString":"OK","values":{"stringList":{"values":["Iris_rtlm-out-neil.csv:s3","Iris_rtlm.csv:s3","BigData11g_num.csv:s3","Smaato_Bids_20130812_CTR.csv:s3"]}}}

Delete Data Source

This API function deletes a data source.

URL

DELETE /symetry/rest/{cid}/dss/{dsname}

HTTP Responses

HTTP Status CodeHTTP Status MessageDescription
200OKSuccess.
409CONFLICTData source cannot be deleted. A SymetryML project might be using the data source. The response contains an error string with further details.

HTTP Response Entity

None

Data Source Information

This API function returns information about a data source. Because this information can be sensitive, the server encrypts it to ensure that the information will be returned safely, even when using HTTP. Use your SymetryML secret key to decrypt the information. For information about how to decrypt the response, see the section Data Source Encryption.

URL

GET /symetry/rest/{cid}/dss/{dsname}

HTTP Responses

HTTP Status CodeHTTP Status MessageDescription
200OKSuccess.

HTTP Response Entity

HTTP Response EntityExample
DSInfo (encrypted)DSInfo encrypted using the customer secret key.

Sample Request/Response

Request:
GET url="http://charm:8080/symetry/rest/c1/dss/BigData11g_num.csv"

Response:
{"statusCode":"OK","statusString":"OK","values":{"dsinfo":"emUmJ7LXKaICPww/dKRoMRNJEMepL37Eq9CgZfhZPWj93mo3A+C8ucfIOGaPwwn2dip/JEuLFjUT/fjHjy18XKFnzFz5Ujp0WmS0uA4ssvAJwNPL6BvnsY6+a/lKa+c/q9/5tz5lr13N13I7OGAhuYhXYV+xb8oFZqsn+bH5spBXRb5u+oyEMXNKLCaNt3pzc/xCyW47KCwIi9V5iSA+fcJAWfetm9ZsIHNbI6utkxKrqrU5OfLmgriGAP++yQtutlGR7r/bKV1bRc8UDsgsXQg1HgoxHcKXCgsAFDFzqsJmZ/5/uQDc0ytc5Fk85GUx"}}

Data Source Browsing

This API function lists content about a remote data source directory.

URL

POST /symetry/rest/{cid}/projects/dsbrowse/ [Body=DSListingRequest (encrypted)]

HTTP Responses

HTTP Status CodeHTTP Status MessageDescription
200OKSuccess.

HTTP Response Entity

HTTP Response EntityDescription
DSListingResponseContains listing information about the requested directory or folder.

Sample Request/Response

Request:
POST url="http://charm:8080/symetry/rest/c1/projects/dsbrowse"

Body:
drhSjndw6G15pgevCsDqaSfjX9x3hMo+dNqd/MV943Dsd2rl2guhvq2qUhjEORcfKAEjHaoRZMKmSbQB6bcca2YT6HmUyRxuOG0wiKgGy0MOEq7+iIncbX4orpGr4rhro1Frw909Uy8qcWskaInQHJT4EGRPcwxvwInFlea39hsMkycFK4pKlTpanOUYgcv7

Response:
{"statusCode":"OK","statusString":"OK","values":{"dsdirectoryListing":{"ok":true,"dirs":["data2/","folder-a/","folder-datasets/","folder-demo/","folder-dev/","folder-dev-pub-http/","folder-dev-pub-https/","folder-docs/","folder-perf-all-reports/","folder-source/"]}}}

Fetching Sample Data 1

This API function fetches a data source sample by specifying the data source information as part of the request body. The response returns up to 128 lines. This REST call needs a request body that contains an encrypted DSInfo data structure (see the section Data Source Encryption).

URL

POST /symetry/rest/{cid}/dss/sample/preview [body=DSInfo encrypted]

HTTP Responses

HTTP Status CodeHTTP Status MessageDescription
200OKSuccess.

HTTP Response Entity

HTTP Response EntityDescription
DataFrameContains a sample of the data source up to 128 lines.

Sample Request/Response

Request:
POST url="http://charm:8080/symetry/rest/c1/dss/sample/preview

Body:
WTqUHBoXHbl+cMMjdc0zgjBP8e44G1os15V+I4GZgDOr1dX9uOfvY5uK9ZgC9yral9XC1ohD1W+UvkPlKR4dQT00EgCdS2UPgZz2NwwooHOM+KY1Ysf5qZlkFKiOkxwoWH/mr3mvvgdTUpZS8zrDJk3gwsavFT5fe0J2lTR33F1OH7FwxP4qs5nzRbVz546l

Response:
{"statusCode":"OK","statusString":"OK","values":{"dataframe":{"attributeNames":["adexchange","imp_width","imp_height","imp_btype","preference","pub","domain","site_base_url","category","device_ip","device_country","device_dma","device_state","device_city","zip","device_carrier","language","device_os","device_make","device_model","device_osv","lat","lon","restriction_bcat","restriction_badv","position","gender","user_keyword","user_yob","user_age","view_count","campaign","creative","creative_type","winner","bid_price","win_price","clicked","rtlm_ctr_score","rtdm","req_hour","req_day","category_count","bcat_count","badv_count"],"data":[["Smaato","320","50","","APP","New IT Solutions Ltd","","","Technology \u0026 Computing","157.55.32.83","US","","WASHINGTON","REDMOND","98052","","","Unknown","","","","47.67398834","-122.1215134","IAB7-28|IAB19-30|IAB22-1|IAB19-3|IAB17-18|IAB26|IAB25|IAB24|IAB9-9|IAB7-

(...)

,"attributeTypes":["S","C","C","L","S","L","S","S","S","S","S","S","S","S","S","S","S","S","S","S","S","S","S","L","S","S","S","S","S","S","B","S","S","S","B","C","C","B","S","B","C","C","B","C","B"]}}}

Fetching Sample Data 2

URL

GET /symetry/rest/{cid}/dss/{dsname}/sample

HTTP Responses

HTTP Status CodeHTTP Status MessageDescription
200OKSuccess.

HTTP Response Entity

HTTP Response EntityDescription
DataFrameSee DataFrame.

Sample Request/Response

Request:
GET url="http://charm:8080/symetry/rest/c1/dss/Smaato_Bids_20130812_CTR.csv/sample"

Response:
{"statusCode":"OK","statusString":"OK","values":{"dataframe":{"attributeNames":["adexchange","imp_width","imp_height","imp_btype","preference","pub","domain","site_base_url","category","device_ip","device_country","device_dma","device_state","device_city","zip","device_carrier","language","device_os","device_make","device_model","device_osv","lat","lon","restriction_bcat","restriction_badv","position","gender","user_keyword","user_yob","user_age","view_count","campaign","creative","creative_type","winner","bid_price","win_price","clicked","rtlm_ctr_score","rtdm","req_hour","req_day","category_count","bcat_count","badv_count"],"data":[["Smaato","320","50","","APP","New IT Solutions Ltd","","","Technology \u0026 Computing","157.55.32.83","US","","WASHINGTON","REDMOND","98052","","","Unknown","","","","47.67398834","-122.1215134","IAB7-28|IAB19-30|IAB22-1|IAB19-3|IAB17-18|IAB26|IAB25|IAB24|IAB9-9|IAB7-

(...)

,"attributeTypes":["S","C","C","L","S","L","S","S","S","S","S","S","S","S","S","S","S","S","S","S","S","S","S","L","S","S","S","S","S","S","B","S","S","S","B","C","C","B","S","B","C","C","B","C","B"]}}}

SymetryML Project Data Source API

Add Data Source to a SymetryML project

This API function lets you add a data source to a project.

URL

GET /symetry/rest/{cid}/projects/{pid}/addds/{dsname}

HTTP Responses

HTTP Status CodeHTTP Status MessageDescription
200OKSuccess.

HTTP Response Entity

None

Remove Data Source from a SymetryML project

This API function lets you remove a data source from a project.

URL

GET /symetry/rest/{cid}/projects/{pid}/detachds/{dsname}

HTTP Responses

HTTP Status CodeHTTP Status MessageDescription
200OKSuccess.

HTTP Response Entity

None

Learning Data from a Data Source

This API function lets you learn from a previously created data source.

URL

GET /symetry/rest/{cid}/projects/{pid}/dss/{dsname}/learn

HTTP Responses

HTTP Status CodeHTTP Status MessageDescription
200OKSuccess. Includes an HTTP Location header specifying the location of the job ID that was created to handle the request. Example: {"statusCode":"ACCEPTED","statusString":"Job Created","values":{}}

HTTP Response Entity

None

Sample Request/Response

Request 1 create ds:
POST url=http://charm:8080/symetry/rest/c1/dss/create

Body 1:
{ENCRYPTED}

Request 2 learn ds:
POST url=http://charm:8080/symetry/rest/c1/projects/r1/dss/irisds/learn

BODY 2
{"errorHandling":1,"attributeNames":["sepal_length","sepal_width","petal_length","petal_width","sepal_lengt_b1","sepal_lengt_b2","sepal_width_b1","sepal_width_b2","petal_length_b1","petal_length_b2","petal_width_b1","petal_width_b2","Iris_setosa","Iris_versicolor","Iris_virginica"],"data":[],"attributeTypes":["C","C","C","C","B","B","B","B","B","B","B","B","B","B","B"]}

Response 2:
{"statusCode":"ACCEPTED","statusString":"Job Created","values":{}}

Response 2 Header:
Location: http://charm:8080/symetry/rest/c1/jobs/4

Job 3 Request:
GET url="http://charm:8080/symetry/rest/c1/jobs/4

Job 3 Response:
{"statusCode":"OK","statusString":"Job is finished","values":{"smlInfo":{"pid":"r1","isDirty":true,"modelsList":[],"modelTypeList":[],"attributeNames":["sepal_length","sepal_width","petal_length","petal_width","sepal_lengt_b1","sepal_lengt_b2","sepal_width_b1","sepal_width_b2","petal_length_b1","petal_length_b2","petal_width_b1","petal_width_b2","Iris_setosa","Iris_versicolor","Iris_virginica"],"attributeIndexes":[0,1,2,3,4,5,6,7,8,9,10,11,12,13,14],"attributeTypes":["C","C","C","C","B","B","B","B","B","B","B","B","B","B","B"],"modelAssessments":{},"modelPredictions":{},"hash":-1,"categorySeparator":"$","type":"cpu","creationDate":1488220762857,"lastModificationDate":1488226030143,"loaded":true,"persisted":true}}}

Forgetting Data from a Data Source

This API function lets you forget data from a previously created data source.

URL

GET /symetry/rest/{cid}/projects/{pid}/dss/{dsname}/forget

HTTP Responses

HTTP Status CodeHTTP Status MessageDescription
200OKSuccess. Includes an HTTP Location header specifying the location of the job ID that was created to handle the request. For example: {"statusCode":"ACCEPTED","statusString":"Job Created","values":{}}

HTTP Response Entity

None

Sample Request/Response

See the section previous Learn endpoint and replace learn with forget in the url.

Prediction Based on a Data Source

After a model is built, you can use this API function to make predictions using a data source. This action can be performed on very large files if they reside on Amazon S3. A prediction file is created that contains the original rows, along with additional prediction information based on the type of model used.

URL

Request:
POST /symetry/rest/{cid}/projects/{pid}/dss/predict/{modelid}?indsname={indsname}&outdsname={outdsname}&impute=false
Request Body:
{
  "attributeNames":[{input attributes names}],
  "attributeTypes":[{input attribute types}]
}

Query Parameters

ParameterRequired / OptionalDescription
indsnameRequiredData source to use as input file for prediction.
outdsnameRequiredData source to use as output file for prediction.
imputeOptionalBoolean parameter specifying whether to impute missing values.

HTTP Responses

HTTP Status CodeHTTP Status MessageDescription
202ACCEPTEDSuccess. Includes an HTTP Location header specifying the location of the job ID that was created to handle the request.
500INTERNAL SERVER ERRORIf the server refuses to accept the new job, it notifies the client with the error "Job execution was refused by server."

HTTP Response Entity

None

Sample Request/Response

Request:
POST url="http://charm:8080/symetry/rest/c1/projects/irisTest/dss/predict/testLDA?indsname=dsin&outdsname=dsout&impute=false
Request Body:
{
  "attributeNames":["sepal_length","sepal_width","petal_length","petal_width"],
  "attributeTypes":["C","C","C","C"]
}

Response Header:
Location: http://charm:8080/symetry/rest/c1/jobs/2

Response:
{"statusCode":"ACCEPTED","statusString":"Job Created","values":{}}

Job Request:
GET url="http://charm:8080/symetry/rest/c1/jobs/2

Job Response:
{"statusCode":"OK","statusString":"Job is finished","values":{"dataframe":{"attributeNames":["resZ_Result","res_Result","normZ_Result","sepal_length","sepal_width","petal_length","petal_width","sepal_lengt_b1","sepal_lengt_b2","sepal_width_b1","sepal_width_b2","petal_length_b1","petal_length_b2","petal_width_b1","petal_width_b2","Iris_setosa","Iris_versicolor","Iris_virginica"],"data":[["-0.4559922893809558","1","0.2929126883636288","4.3","3","1.1","0.1","1","0","0","1","1","0","1","0","1","0","0"],["-0.5225734165408404","1","0.2965158290613048","

(…)

["-2.6526076600477295","1","0.4232667751179995","6.9","3.1","5.4","2.1","0","1","0","1","0","1","0","1","0","0","1"]]}}}

Encoder Data Source API

Updating an Encoder with a Data Source

This API function updates an Encoder with data from a data source.

URL

GET /symetry/rest/{cid}/encoders/{encodername}/learnds [body = DSInfo (encrypted)]

HTTP Responses

HTTP Status CodeHTTP Status MessageDescription
202ACCEPTEDSuccess. Includes an HTTP Location header specifying the location of the job ID that was created to handle the request.
500INTERNAL SERVER ERRORIf the server refuses to accept the new job, it notifies the client with the error "Job execution was refused by server."

HTTP Response Entity

None

Sample Request/Response

Request:
POST url="http://charm:8080/symetry/rest/c1/encoders/enctest/learnds

Request Body:
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

Response Header:
Location: http://charm:8080/symetry/rest/c1/jobs/4

Response:
{"statusCode":"ACCEPTED","statusString":"Job Created","values":{}}

Job Request:
GET url="http://charm:8080/symetry/rest/c1/jobs/4

Job Response:
{"statusCode":"OK","statusString":"Job is finished","values":{}}

Listing a Data Source Used by an Encoder

This API function lists the data source(s) that were used to update an Encoder.

URL

GET /symetry/rest/{cid}/encoders/{encodername}/dss

HTTP Responses

HTTP Status CodeHTTP Status MessageDescription
202ACCEPTEDSuccess. Includes an HTTP Location header specifying the location of the job ID that was created to handle the request.

HTTP Response Entity

HTTP Response EntityExample
StringListSee StringList

Sample Request/Response

Request:
GET url="http://charm:8080/symetry/rest/c1/encoders/enctest/dss

Response:
{"statusCode":"OK","statusString":"OK","values":{"stringList":{"values":["s12"]}}}

Data Source Job Status

When invoking a JobStatus for a job that was initiated for learning, forgetting, or making predictions based on a data source, the response might contain an entity. The following sections describe these cases.

Learning a Data Source

HTTP Responses

HTTP Status CodeHTTP Status MessageDescription
200OKNo entity.
202ACCEPTEDJob is not finished. Includes a JobInfo entity. Using the current and maximum fields, you can compute the percentage of the job that was accomplished so far. "Maximum" refers to the size of the file to process and contains the approximate number of bytes processed by the job.

Forgetting a Data Source

HTTP Responses

HTTP Status CodeHTTP Status MessageDescription
200OKNo entity.
202ACCEPTEDJob is not finished. Includes a JobInfo entity. For more information, see the section JobInfo and also the section on Symetry Jobs. Using the current and maximum fields, you can compute the percentage of the job that was accomplished so far. "Maximum" refers to the size of the file to process and contains the approximate number of bytes processed by the job.

Prediction Based on Data Source

HTTP Responses

HTTP Status CodeHTTP Status MessageDescription
200OKDataFrame that contains a sample of the predictions (up to 128 lines). Because Amazon files can be very large, it is not possible to return the prediction result file in its entirety within a REST call. Use your favorite tool to fetch the prediction results from the data source (S3 or SFTP). The prediction result file contains all the original file columns, plus the additional prediction column for each row. Any additional columns depend on the type of model used to make the predictions.
202ACCEPTEDJob is not finished. Includes a JobInfo entity. Using the current and maximum fields, it is possible to compute the percentage of the job that was accomplished so far. "Maximum" refers to the size of the file to process and contains the approximate number of bytes processed by the job.

On this page

Using a Data SourceRequired DSInfo FieldsAdditional Information Stored in Data SourceHTTP/HTTPS Data SourceSecure FTP (SFTP)Amazon S3Oracle OCI Object Storage with S3 CompatibilityGoogle Cloud StorageMicrosoft Azure Blob StorageAmazon RedshiftSpark Map ReduceJDBCSnowflakeAmazon EMRAdditional CSV OptionsAdditional Information on Spark S3 Data SourceAbout Data Source Plugins (DSPlugins)Data Source EncryptionData Source CreateURLHTTP ResponsesHTTP Response EntitySample Request/ResponseData Source UpdateURLHTTP ResponsesHTTP Response EntitySample Request/ResponseList Customer Data SourcesURLHTTP ResponsesHTTP Response EntitySample Request/ResponseDelete Data SourceURLHTTP ResponsesHTTP Response EntityData Source InformationURLHTTP ResponsesHTTP Response EntitySample Request/ResponseData Source BrowsingURLHTTP ResponsesHTTP Response EntitySample Request/ResponseFetching Sample Data 1URLHTTP ResponsesHTTP Response EntitySample Request/ResponseFetching Sample Data 2URLHTTP ResponsesHTTP Response EntitySample Request/ResponseSymetryML Project Data Source APIAdd Data Source to a SymetryML projectURLHTTP ResponsesHTTP Response EntityRemove Data Source from a SymetryML projectURLHTTP ResponsesHTTP Response EntityLearning Data from a Data SourceURLHTTP ResponsesHTTP Response EntitySample Request/ResponseForgetting Data from a Data SourceURLHTTP ResponsesHTTP Response EntitySample Request/ResponsePrediction Based on a Data SourceURLQuery ParametersHTTP ResponsesHTTP Response EntitySample Request/ResponseEncoder Data Source APIUpdating an Encoder with a Data SourceURLHTTP ResponsesHTTP Response EntitySample Request/ResponseListing a Data Source Used by an EncoderURLHTTP ResponsesHTTP Response EntitySample Request/ResponseData Source Job StatusLearning a Data SourceHTTP ResponsesForgetting a Data SourceHTTP ResponsesPrediction Based on Data SourceHTTP Responses