Add details to format exceptions (#210)

We do not expect format exceptions to occur unless the server API is
changing. Change from a plain `FormatException` to a specific exception
which indicates the two possible resolution strategies - upgrade if not
already on the latest SDK, or file an issue.

Default the role to 'model' for chat responses (#205)

Fixes #197

A content requires a role when it is sent to the model in the history.
If the backend happens to respond with a message that has no role,
default it to 'model'.

Allow Vertex formatted CitationMetadata (#199)

Fixes #184

The vertex backend uses a similar data model, but names the field
`citations` instead of `citationSources`. Read either field, whichever
exists, and parse it otherwise consistently.

Add a request escape hatch for the vertex SDK (#180)

Maintains the internal details of using the client, but offers a general
escape hatch for any unary request by exposing `Task`.

Add a `makeRequest` utility method in an extension, and update the
existing unary requests to use it.

Prepare to publish (#176)

Prepare to publish (#172)

Add Content.functionResponses utility (#159)

The alternative when replying to multiple functions in parallel is to
use the `Content` constructor and pass the string `'function'` to the
`role` parameter. Use a static method to obscure the role string for
consistency with the others.

Update sample with a demonstration of replying to all function calls.

Add usageMetadata to GenerateContentResponse (#143)

Add `UsageMetadata` and relevante parsing. Add a `usageMetadata` field
on `GenerateContentResponse`.

Add usage of the new field to the advanced text sample.

Refactor the `GenerateContentResponse` parse method to handled each
field individually at the top level. There is a behavior change for an
error case, but it is not visible through the message formats that are
returned from the backend in practice.

Prepare to publish.

Fix typo in function_calling.dart (#132)

Prepare to publish (#126)