Excerpt |
---|
This section shows all the available options for the Data Loader and explains their usage. |
...
The purpose of the Data Loader tool is to extract data from different sources: supported supported databases, CSV and excel Excel files (xls and xlsx) and then upload them into Squirro.
...
Argument | Mandatory | Description | ||
---|---|---|---|---|
General Options | ||||
--help, -h | Show a help message and exit. | |||
--version | Output the tool version and exit. | |||
--verbose, -v | Increase log verbosity.
| |||
--log-file | Path to a log file on disk, where the log output is to be stored. If this is not specified, the log messages are shown on the console. | |||
--parallel-uploaders NUMBER | Number of uploaders (default is 1).
| |||
--meta-db-dir PATH | Directory of the SQLite metadata database. | |||
--meta-db-file STRING | File name of the SQLite metadata database. | |||
--transform-items | Apply item transformation on the client (default is false). This flag must be used when loading data to Squirro servers < v3.3.1, or optionally when for some reason you would like to perform the item transformation on your local machine. By default (Squirro >= v3.3.1) the item transformation takes place on the server, and specifically in the Pipeline Workflow by the Transform Input step. | |||
Connection Options (see Connecting to Squirro for finding these values) | ||||
--token TOKEN | Yes | The Authentication Token with which to authenticate. If the token value starts with a dash, you need to use an equal sign to specify the value like this: --token="-12345…" | ||
--cluster URL | The Squirro Cluster into which to import the data. | |||
--project-id PROJECT_ID | Yes | The Project identifier into which to import the data. | ||
Testing Options | ||||
--limit LIMIT | If set, then only this many items are sent to Squirro. This can be used to test the options on a small subset of the data to make sure the mapping, facets and pipelets all work correctly. | |||
--dry-run, -n | Do all the processing, except server-side actions (no uploads or facet configuration). | |||
Source - Item Mapping Options | ||||
--map-title STRING | Which column is mapped to the "title" field. | |||
--map-abstract STRING | Which column is mapped to the "summary" field. | |||
--map-created-at STRING | Which column is mapped to the "created_at" field. | |||
--map-id STRING | Which column is mapped to the "external_id" field. | |||
--map-body [STRING…] | Which columns are mapped to the "body" field. | |||
--map-body-mime STRING | Which column is mapped to the MIME type of the body. Use The typical values in this field for this are either | |||
--body-mime STRING | Fixed MIME type of the body. Use The typical values for this are either | |||
--map-url STRING | Which column is mapped to the "link" field. | |||
--map–webshot-url STRING | Which column is mapped to the webshot_url for thumbnail extraction | |||
--map–webshot-picture-hint STRING | Which column is mapped to the webshot_picture_hint. | |||
--map-file-name STRING | Which column is mapped to the file-name. | |||
--map-file-mime STRING | Which column is mapped to file mime type. | |||
--map-file-data STRING | Which column is mapped to file contents. | |||
--map-file-compressed STRING | Which column specifies if the file is compressed with gz. Possible values should be 'y, yes, t, true, 1', case insensitive. | |||
--map-flag STRING | Which column determines if the received record is an insert/update or delete. If the value is 'd' the record is deleted, otherwise is considered an insert/update | |||
--map-language STRING | Which column is matched to the language. | |||
ItemUploader Options (see ItemUploader documentation for more information) | ||||
--object-id OBJECT_ID | Object identifier. | |||
--source-id SOURCE_ID | Source identifier, defaults to the input file name. | |||
--source-name SOURCE_NAME | Source name, defaults to the input file name. | |||
--enable-filtering | Enable item filtering to support alerts. (deprecated, use a suitable pipeline workflow) | |||
--enable-near-duplicate-detection | Enable near duplicate detection to link similar documents in the Squirro user interface. (deprecated, use a suitable pipeline workflow) | |||
--batch-size NUMBER | Batch size for uploads (default is auto - change value based on the size of the payload). | |||
Item pre-processing Options | ||||
--body-template-file PATH | Jinja2 html template file with full path. | |||
--title-template-file PATH | Jinja2 html template file with full path. | |||
--abstract-template-file PATH | Jinja2 html template file with full path. | |||
--pipelets-file PATH | JSON file containing the pipelets called by the db loader in execution order. | |||
--pipelets-error-behavior STRING | Specify job behavior in case a pipelet raises an exception. Valid values are | |||
--facets-file PATH | JSON file containing facets configuration. | |||
Source Options | ||||
--pipeline-workflow-name | Select the pipeline workflow by name. Only interpreted if --pipeline-workflow-id is not set. (introduced in Squirro 2.6.0) | |||
--pipeline-workflow-id | Select the pipeline workflow by ID. If not set, use default workflow. (introduced in Squirro 2.6.0) | |||
--source-type STRING | Type of source to load data from. Valid values are | |||
--source-script PATH | Path of the Data Loader Plugin Python script. | |||
--source-batch-size NUMBER | Batch size for source unloads (default is "1000"). | |||
--incremental-column STRING | Which column will be used as incremental reference - usually for a datetime column. If missing a full load will be done. | |||
--job-id STRING | Used for job locking and storing the last-known value of incremental columns. If not given, this is calculated based on the source parameters. | |||
--reset | Deletes incremental date information for the current sql query. Useful to perform an incremental load with reset. |
...
Code Block | ||
---|---|---|
| ||
squirro_data_load -v \ --token $token \ --cluster $cluster \ --project-id $project_id \ --source-name csv_sample \ --source-file sample.csv \ --source-type csv \ --csv-delimiter , \ --csv-quotechar " \ --map-title Title \ --map-id ID \ --map-body Description |
Note that the lines have been wrapped with the backslash (\) at the end of each line. On bash/windows setup you will need to use circumflex (^) instead.
This example assumes that $token, $cluster and $project_id have been declared beforehand.
Excel Source Options
When using an excel file as the data source, only full load is supported and data must have a header in order to determine the schema. If the first row of the data (after applying the boundaries, if needed) is not the header, a KeyError exception will be raised and the job will stop. In this case it’s not possible to determine the schema of the data.
The command line parameters used for an excel data source:
Argument | Mandatory | Description |
---|---|---|
--excel-sheet STRING | Excel sheet name. Default: get first sheet. | |
--excel-boundaries NUMBER: NUMBER | Limit rows loaded from excel. Format is: start_row:rows_discarded_from_end. | |
--source-file PATH | Yes | Path of excel data file. |
Usage
The example below shows a simple load from an excel file, mapping only the title, id and body of the Squirro item to columns from the ‘sample.xlsx’ excel file, without using any of the additional files for facets, templating, piplets etc. The Data Loader tool will only load the ‘Products’ sheet of the file and from this sheet the rows starting at 1 up to the last 100 rows, which will not be loaded.
Code Block | ||
---|---|---|
| ||
squirro_data_load -v \ --token $token \ --cluster $cluster \ --project-id $project_id \ --source-name excel_sample \ --source-file sample.xlsx \ --source-type excel \ --excel-sheet Products \ --excel-boundaries 1:100 \ --map-title Title \ --map-id ID \ --map-body Description |
Note that the lines have been wrapped with the backslash (\) at the end of each line. On bash/windows setup you will need to use circumflex (^) instead.
This example assumes that $token, $cluster and $project_id have been declared beforehand.
JSON Source Options
When using an JSON file as the data source, only full load is supported.
Schema of the data is determined using the first item found, therefore assuming that all items have the same structure. If that's not the case, the loader might fail.
The command line parameters used for a JSON data source:
Argument | Mandatory | Description |
---|---|---|
--item-schema STRING | No | In case the JSON objects are not available as a top level structure, use this parameter to un-nest the JSON structure |
--source-file PATH | No, one of --source-file or --source-folder must be provided | Path of JSON data file. |
--source-folder PATH | No, one of --source-file or --source-folder must be provided | Path of directory containing multiple JSON files. Only available in CLI mode |
Usage
The example below shows a load from a nested JSON file (see attachment.json for an example and facets.json to demonstrate how to map facets)
Code Block | ||
---|---|---|
| ||
squirro_data_load -vv \ --cluster "$CLUSTER" \ --token "$TOKEN" \ --project-id "$PROJECT_ID" \ --source-type 'json' \ --map-title 'data.headline' \ --map-body 'data.body' \ --map-id 'data.id' \ --map-created-at 'data.versionCreated' \ --source-name 'JSON WIKI TEST' \ --facets-file 'facets.json' \ --source-file "$SOURCE_FILE" \ --item-schema 'Items' |
Note that the lines have been wrapped with the backslash (\) at the end of each line. On bash/windows setup you will need to use circumflex (^) instead.
This example assumes that $TOKEN, $CLUSTER and $PROJECT_ID have been declared beforehand.
Database Options
When loading from a database, both full and incremental load are supported, using a select query supplied as a string or in a file. The script uses uses SQLAlchemy to connect to any database.
Tested databases:
- Postgres and all databases using the postgres driver for connection (Greenplum, Redshift etc)
- Microsoft SQL
- Oracle
- MySQL
- SQLite
The command line parameters used for a database source:
Argument | Mandatory | Description |
---|---|---|
--db-connection STRING | Yes | Database connection string. |
--input-file PATH | File containing the SQL code. | |
--input-query STRING | SQL query. |
Note that the --input-file
and --input-query
arguments are mutually exclusive.
Usage
In the following example we are performing a simple load from the database, mapping the title, id and body of the Squirro item to columns from a database table that is interrogated in the sample.sql file. The Data Loader tool makes a full load of all the rows specified in the sample.sql file since the argument --incremental-column
is not set.
Code Block | ||
---|---|---|
| ||
squirro_data_load -v \ --token $token \ --cluster $cluster \ --project-id $project_id \ --db-connection $db_connection_string \ --source-name db_sample \ --input-file $script_dir/interaction.sql \ --source-type database \ --map-title Title \ --map-id ID \ --map-body Description |
Note that the lines have been wrapped with the backslash (\) at the end of each line. On bash/windows setup you will need to use circumflex (^) instead.
This example assumes that $token, $cluster and $project_id have been declared beforehand.
...
Option | Mandatory | Description | ||
---|---|---|---|---|
--folder PATH | No, one of the --folder or --zip-file-path must be provided | Filesystem location that will be indexed in Squirro | ||
--zip-file-path | No, one of the --folder or --zip-file-path must be provided | Absolute path to a zip file containing all the files " "to be imported into Squirro | ||
--deletions | No | If set, then any files that are no longer present on the file system are also removed from Squirro. To use this, the
| ||
--include-file PATH | No | Path to a file containing inclusion rules. This is a list of patterns that files need to be match to be indexed. If provided, then only files that match at least one pattern are indexed. | ||
--exclude-file PATH | No | Path to a file containing exclusion rules. This is a list of patterns for files that should not be indexed. Any file that matches at least one such pattern is not indexed, independent of whether it also matches the include rules. | ||
--skip-errors | No | Ignore any file system errors when processing individual files. This way a single file system read error does not prevent the entire load from succeeding. If the error is temporary, then the file will be picked up in the next load. | ||
Performance Optimisations | ||||
--convert-file PATH | No | Path to a file containing conversion file patterns. Files that match any of these rules will be indexed with full content. See Content Conversion for the file types that Squirro supports full indexing for. By limiting to a smaller number of extensions, this allows the file system loader to only process content in Squirro for which indexing will be effective. | ||
--file-size-limit | No | Maximum size in megabytes of files that should be indexed with content. Also see --index-all below. | ||
--index-all | No | If set, then files over the --file-size-limit are indexed, but without their content. In the default case of this not being set, those files are skipped entirely. | ||
--batch-size-limit | No | Maximum size of requests sent to Squirro's API for indexing of files. | ||
--deduplicate | No | Deduplicate files based on file content. Exact duplicates are only ever indexed ones, with duplicates ignored. | ||
Logging and Debugging | ||||
--log-excludes | No | Log matches for inclusion/exclusion rules. | ||
--progress | No | Log progress verbosely. |
...
Code Block | ||
---|---|---|
| ||
squirro_data_load -v \ --token $TOKEN \ --cluster $CLUSTER \ --project-id $PROJECT_ID \ --source-type filesystem \ --folder FOLDER_TO_INDEX \ --map-title "title" \ --map-file-name "file_name" \ --map-file-mime "file_mime" \ --map-file-data "file_data" \ --map-id "id" \ --map-url "link" \ --map-created-at "created_at" \ --facets-file facets.json |
Note that the lines have been wrapped with the backslash (\) at the end of each line. On bash/windows setup you will need to use circumflex (^) instead.
This example assumes that $token, $cluster and $project_id have been declared beforehand.
Squirro Options
When Loading data from any squirro source following command line parameters can be used:
...
Code Block | ||
---|---|---|
| ||
squirro_data_load -v \ --token $TOKEN \ --cluster $CLUSTER \ --project-id $PROJECT_ID \ --source-cluster $SOURCE_CLUSTER \ --source-token $SOURCE_TOKEN \ --source-project-id $SOURCE_PROJECT_ID \ --source-type squirro \ --source-query "*" \ --include-facets \ --include-entities \ --map-title "title" \ --map-id "id" \ --map-url "link" \ --map-created-at "created_at" \ --progress \ --deduplicate \ --retry 5 |
Note that the lines have been wrapped with the backslash (\) at the end of each line. On bash/windows setup you will need to use circumflex (^) instead.
This example assumes that $token, $cluster and $project_id have been declared beforehand.
Feed Options
When Loading data from any feed source following command line parameters can be used:
Argument | Mandatory | Description |
---|---|---|
--feed-sources | Yes | Space-separated list of URLs (strings). |
--query-timeout | No | Timeout (in seconds) for fetching the feed |
--max-backoff | No | Maximum number of hours to wait if the feed update frequency is low |
--custom-date-field | No | For non-standard rss datetime fields, enter the field |
--custom-date-format | No | For non-standard rss datetime formats, enter the format i.e. %m/%d/%Y. |
--rss-username | No | Username for RSS Basic Authentication |
--rss-password | No | Password for RSS Basic Authentication |
Usage
A simple example for loading data from feed source is given by:
Code Block | ||
---|---|---|
| ||
squirro_data_load -v \ --token $TOKEN \ --cluster $CLUSTER \ --project-id $PROJECT_ID \ --source-type feed \ --source-name feed_sample \ --feed-sources 'https://www.theregister.co.uk/headlines.atom' 'http://rss.nytimes.com/services/xml/rss/nyt/HomePage.xml' \ --map-title "title" \ --map-id "id" \ --map-body "description" \ --map-created-at "created_at" \ --batch-size 100 \ --source-batch-size 100 \ --facets-file facets.json |
Note that the lines have been wrapped with the circumflex (^) at the end of each line. On Mac and Linux you will need to use backslash (\) instead.
User defined sources
If data needs to be extracted from other sources than the ones described above there is the option to write a custom source.
To do this a new Python module must be created and has to implement the abstract base class DataSource.
In this way the Data Loader can index data from other sources without modifications.
...