Simplify and generalize CI scripts

[MichaelKohler]

Right now the Ci scripts are quite specific to Wikipedia extracts. We can generalize this and make it possible to also use for other targets.

Ideas:

• Split up tasks into separate scripts files
• Pass actual cargo command to script instead of assuming extract
• General cleanup to make it easier to understand the different flows (PR, sample extract, blocklist, ..)

[MichaelKohler]

/action blocklist en 80

[github-actions]

Job started: https://github.com/common-voice/cv-sentence-extractor/actions/runs/888614010

[MichaelKohler]

/action blocklist az 80

[github-actions]

Job started: https://github.com/common-voice/cv-sentence-extractor/actions/runs/888783585

[github-actions]

Job finished: https://github.com/common-voice/cv-sentence-extractor/actions/runs/888783585
Don't forget to download the artifacts.