xvc pipeline import

Synopsis

$ xvc pipeline import --help
Import the pipeline from a file

Usage: xvc pipeline import [OPTIONS]

Options:
  -p, --pipeline-name <PIPELINE_NAME>  Name of the pipeline to import. If not set, the name from the file is used
      --file <FILE>                    File to read the pipeline. Use stdin if not specified
      --format <FORMAT>                Input format. One of json or yaml. If not set, the format is guessed from the file extension. If the file extension is not set, json is used as default
      --overwrite                      Overwrite the pipeline even if the name already exists
  -h, --help                           Print help

Examples

This command is used to import pipelines exported with xvc pipeline export.

You can edit and import the pipelines exported with the command.

Xvc doesn't guarantee that the format of these files will be compatible across versions. You can use these files to share pipeline definitions but it may not be a good way to store pipeline definitions for longer periods.

This command works only in Xvc repositories.

$ git init
...
$ xvc init

The following file generated with xvc pipeline export.

$ cat pipeline.yaml
version: 1
name: default
workdir: ''
steps:
- name: step1
  command: touch abc.txt
  invalidate: ByDependencies
  dependencies: []
  outputs: []
- name: step2
  command: touch def.txt
  invalidate: ByDependencies
  dependencies:
  - !Step
    name: step1
  - !Generic
    generic_command: ping -c 2 example.com
    output_digest: null
  - !GlobItems
    glob: '*.txt'
    xvc_path_metadata_map: {}
    xvc_path_content_digest_map: {}
  - !Glob
    glob: '*.txt'
    xvc_paths_digest: null
    xvc_metadata_digest: null
    content_digest: null
  - !RegexItems
    path: requirements.txt
    regex: ^tensorflow
    lines: []
    xvc_metadata: null
  - !Regex
    path: requirements.txt
    regex: ^tensorflow
    lines_digest: null
    xvc_metadata: null
  - !Param
    format: YAML
    path: params.yaml
    key: model.conv_units
    value: null
    xvc_metadata: null
  - !LineItems
    path: params.yaml
    begin: 1
    end: 20
    xvc_metadata: null
    lines: []
  - !Lines
    path: params.yaml
    begin: 1
    end: 20
    xvc_metadata: null
    digest: null
  - !UrlDigest
    url: https://example.com/
    etag: null
    last_modified: null
    url_content_digest: null
  outputs:
  - !File
    path: def.txt
  - !Metric
    path: metrics.json
    format: JSON
  - !Image
    path: plots/confusion.png

You can import this file to construct the pipeline at once. Note that the export command outputs JSON by default.

$ xvc pipeline import --file pipeline.yaml --overwrite

$ xvc pipeline export
{
  "name": "default",
  "steps": [
    {
      "command": "touch abc.txt",
      "dependencies": [],
      "invalidate": "ByDependencies",
      "name": "step1",
      "outputs": []
    },
    {
      "command": "touch def.txt",
      "dependencies": [
        {
          "Step": {
            "name": "step1"
          }
        },
        {
          "Generic": {
            "generic_command": "ping -c 2 example.com",
            "output_digest": null
          }
        },
        {
          "GlobItems": {
            "glob": "*.txt",
            "xvc_path_content_digest_map": {},
            "xvc_path_metadata_map": {}
          }
        },
        {
          "Glob": {
            "content_digest": null,
            "glob": "*.txt",
            "xvc_metadata_digest": null,
            "xvc_paths_digest": null
          }
        },
        {
          "RegexItems": {
            "lines": [],
            "path": "requirements.txt",
            "regex": "^tensorflow",
            "xvc_metadata": null
          }
        },
        {
          "Regex": {
            "lines_digest": null,
            "path": "requirements.txt",
            "regex": "^tensorflow",
            "xvc_metadata": null
          }
        },
        {
          "Param": {
            "format": "YAML",
            "key": "model.conv_units",
            "path": "params.yaml",
            "value": null,
            "xvc_metadata": null
          }
        },
        {
          "LineItems": {
            "begin": 1,
            "end": 20,
            "lines": [],
            "path": "params.yaml",
            "xvc_metadata": null
          }
        },
        {
          "Lines": {
            "begin": 1,
            "digest": null,
            "end": 20,
            "path": "params.yaml",
            "xvc_metadata": null
          }
        },
        {
          "UrlDigest": {
            "etag": null,
            "last_modified": null,
            "url": "https://example.com/",
            "url_content_digest": null
          }
        }
      ],
      "invalidate": "ByDependencies",
      "name": "step2",
      "outputs": [
        {
          "File": {
            "path": "def.txt"
          }
        },
        {
          "Metric": {
            "format": "JSON",
            "path": "metrics.json"
          }
        },
        {
          "Image": {
            "path": "plots/confusion.png"
          }
        }
      ]
    }
  ],
  "version": 1,
  "workdir": ""
}

If you don't supply the --overwrite option, Xvc will report an error and quit.

$ xvc pipeline import --file pipeline.yaml
[ERROR] Pipeline Error: Pipeline default already found

You can specify a new name for the pipeline and it will override the name set in the file. This way you can edit and import similar pipelines with minor differences.

$ xvc pipeline import --pipeline-name another-pipeline --file pipeline.yaml

You can also use stdin to import a pipeline but you must specify the input format.