Create a version number for the crawler

[adilsoncarvalho]

Resolves #15

[adilsoncarvalho]

The code added on f6392bf will look for a file within the class file with the .version extension to get the version number. If it is not present it'll be assumed a dev version.

For each spider class the CI must provide a version file with the short sha1 so it will be loaded and used as the version for the spider.

The CI must be smart enough to detect that that spider got changed to it would be important to change its version. Maybe there is a way to get from git the last commit that changed it.

The file structure

# the class file
/scrapy/nfcrawler/nfcrawler/spiders/pr_nfce.py

# the expected version file
/scrapy/nfcrawler/nfcrawler/spiders/pr_nfce.version

The versions

# during development
spider-name/dev

# production
spider-name/f6392bf

[adilsoncarvalho]

Getting the sha1 of the last change on file

git rev-list HEAD -1 -- FILENAME

There is no --short option 😭

➜  barateza-nfcrawler git:(15-issue) git rev-list HEAD -1 -- nfcrawler/nfcrawler/spiders/pr_nfce.py
f6392bfd8dd119cc6f62c6a726988bc9bb754a58

➜  barateza-nfcrawler git:(15-issue) git rev-list HEAD -1 -- Dockerfile
cdd8b270d06c067d77104f3a538c1d96b37f2964

[adilsoncarvalho]

There is a --short option (with another name) 🎊

➜  barateza-nfcrawler git:(15-issue) git rev-list HEAD -1 --abbrev-commit -- Dockerfile
cdd8b27

[adilsoncarvalho]

Added spider version on commit 9afa76e.

{
  "spider": {"version": "pr-nfce/dev"}
}

[adilsoncarvalho]

Maybe we should consider storing the Scrapy stats. See issue #32