Logic around creating scrape chain_data.json seems broken

[pipermerriam]

What is wrong.

$ ethpm scrape
EthPM CLI v0.1.0a0

Traceback (most recent call last):
  File "/home/piper/python-environments/ethpm-cli/bin/ethpm", line 11, in <module>
    load_entry_point('ethpm-cli', 'console_scripts', 'ethpm')()
  File "/home/piper/projects/ethpm-cli/ethpm_cli/main.py", line 84, in main
    scraper(args)
  File "/home/piper/projects/ethpm-cli/ethpm_cli/main.py", line 46, in scraper
    last_scraped_block = scrape(w3, ethpmcli_dir, start_block)
  File "/home/piper/projects/ethpm-cli/ethpm_cli/scraper.py", line 38, in scrape
    initialize_ethpm_dir(ethpm_dir, w3)
  File "/home/piper/projects/ethpm-cli/ethpm_cli/scraper.py", line 112, in initialize_ethpm_dir
    validate_chain_data_store(chain_data_path, w3)
  File "/home/piper/projects/ethpm-cli/ethpm_cli/validation.py", line 71, in validate_chain_data_store
    f"{chain_data_path} does not appear to be a valid EthPM CLI datastore."
ethpm_cli.exceptions.InstallError: /home/piper/.local/share/ethpmcli/chain_data.json does not appear to be a valid EthPM CLI datastore.

I previously got myself a corrupted chain_data.json file after hacking around with the system a bit. After deleting the file, this is what happens when I try to ethpm scrape

How to fix

ethpm-cli/ethpm_cli/scraper.py

Line 110 in 7d56b01

if ethpm_dir.is_dir():

This code assumes that if the directory is present then the file must be present also. Probably worth adding a test that demonstrates this case as well as one where the document is invalid JSON so that the CLI gives good error messaging in all cases.