devlooped / yaml

YamlPeek MSBuild Task

YamlPeek MSBuild Task

Read values from yaml files using JSONPath query expressions.

Usage:

  <YamlPeek ContentPath="[YAML_FILE]" Query="[JSONPath]">
    <Output TaskParameter="Result" PropertyName="Value" />
  </YamlPeek>
  <YamlPeek Content="[YAML]" Query="[JSONPath]">
    <Output TaskParameter="Result" ItemName="Values" />
  </YamlPeek>

Parameters:

Parameter	Description
Content	Optional string parameter. Specifies the YAML input as a string.
ContentPath	Optional ITaskItem parameter. Specifies the YAML input as a file path.
Query	Required string parameter. Specifies the JSONPath expression.
Result	Output ITaskItem[] parameter. Contains the results that are returned by the task.

You can either provide the path to a YAML file via ContentPath or provide the straight YAML content to Content. The Query is a JSONPath expression that is evaluated and returned via the Result task parameter. You can assign the resulting value to either a property (i.e. for a single value) or an item name (i.e. for multiple results).

YAML object properties are automatically projected as item metadata when assigning the resulting value to an item. For example, given the following JSON:

http:
  host: localhost
  port: 80
  ssl: true

You can read the entire http value as an item with each property as a metadata value with:

<YamlPeek ContentPath="host.yaml" Query="$.http">
    <Output TaskParameter="Result" ItemName="Http" />
</YamlPeek>

The Http item will have the following values (if it were declared in MSBuild):

<ItemGroup>
    <Http Include="[item raw json]">
        <host>localhost</host>
        <port>80</port>
        <ssl>true</ssl>
    </Http>
</ItemGroup>

These item metadata values could be read as MSBuild properties as follows, for example:

<PropertyGroup>
    <Host>@(Http -> '%(host)')</Host>
    <Port>@(Http -> '%(port)')</Port>
    <Ssl>@(Http -> '%(ssl)')</Ssl>
</PropertyGroup>

In addition to the explicitly opted in object properties, the entire node is available as raw YAML via the special _ (single underscore) metadata item.