Configure the Speech CLI datastore

  • Article

The Speech CLI can rely on settings in configuration files, which you can refer to using a @ symbol. The Speech CLI saves a new setting in a new ./spx/data subdirectory that is created in the current working directory for the Speech CLI. The Speech CLI first looks for a configuration value in your current working directory, then in the datastore at ./spx/data, and then in other datastores, including a final read-only datastore in the spx binary.

In the Speech CLI quickstart, you used the datastore to save your @key and @region values, so you didn't need to specify them with each spx command. Keep in mind, that you can use configuration files to store your own configuration settings, or even use them to pass URLs or other dynamic content generated at runtime.

For more details about datastore files, including use of default configuration files (@spx.default, @default.config, and @*.default.config for command-specific default settings), enter this command:

spx help advanced setup

nodefaults

The following example clears the @my.defaults configuration file, adds key-value pairs for key and region in the file, and uses the configuration in a call to spx recognize.

spx config @my.defaults --clear
spx config @my.defaults --add key 000072626F6E20697320636F6F6C0000
spx config @my.defaults --add region chinaeast2

spx config @my.defaults

spx recognize --nodefaults @my.defaults --file hello.wav

Dynamic configuration

You can also write dynamic content to a configuration file using the --output option.

For example, the following command creates a custom speech model and stores the URL of the new model in a configuration file. The next command waits until the model at that URL is ready for use before returning.

spx csr model create --name "Example 4" --datasets @my.datasets.txt --output url @my.model.txt
spx csr model status --model @my.model.txt --wait

The following example writes two URLs to the @my.datasets.txt configuration file. In this scenario, --output can include an optional add keyword to create a configuration file or append to the existing one.

spx csr dataset create --name "LM" --kind Language --content https://crbn.us/data.txt --output url @my.datasets.txt
spx csr dataset create --name "AM" --kind Acoustic --content https://crbn.us/audio.zip --output add url @my.datasets.txt

spx config @my.datasets.txt

SPX config add

For readability, flexibility, and convenience, you can use a preset configuration with select output options.

For example, you might have the following requirements for captioning:

  • Recognize from the input file caption.this.mp4.
  • Output WebVTT and SRT captions to the files caption.vtt and caption.srt respectively.
  • Output the offset, duration, resultid, and text of each recognizing event to the file each.result.tsv.

You can create a preset configuration named @caption.defaults as shown here:

spx config @caption.defaults --clear
spx config @caption.defaults --add output.each.recognizing.result.offset=true
spx config @caption.defaults --add output.each.recognizing.result.duration=true
spx config @caption.defaults --add output.each.recognizing.result.resultid=true
spx config @caption.defaults --add output.each.recognizing.result.text=true
spx config @caption.defaults --add output.each.file.name=each.result.tsv
spx config @caption.defaults --add output.srt.file.name=caption.srt
spx config @caption.defaults --add output.vtt.file.name=caption.vtt

The settings are saved to the current directory in a file named caption.defaults. Here are the file contents:

output.each.recognizing.result.offset=true
output.each.recognizing.result.duration=true
output.each.recognizing.result.resultid=true
output.each.recognizing.result.text=true
output.all.file.name=output.result.tsv
output.each.file.name=each.result.tsv
output.srt.file.name=caption.srt
output.vtt.file.name=caption.vtt

Then, to generate captions, you can run this command that imports settings from the @caption.defaults preset configuration:

spx recognize --file caption.this.mp4 --format any --output vtt --output srt @caption.defaults

Using the preset configuration as shown previously is similar to running the following command:

spx recognize --file caption.this.mp4 --format any --output vtt file caption.vtt --output srt file caption.srt --output each file each.result.tsv --output all file output.result.tsv --output each recognizer recognizing result offset --output each recognizer recognizing duration --output each recognizer recognizing result resultid --output each recognizer recognizing text

Next steps