Encode-utils

Bug fixes in October:


1. Updated ``typecast()`` function in ``eu_register.py`` so that it can handle JSON Schema numbers. Moreover, when an attribute is declared to be a number, this function will convert from string to either int or float based on the given string representation. Thanks for khine for reporting.
1. Fixed bug reported by weiwei where in profile.py the ``Profile.required_properties()`` method always expected the given profile to contain a top-level 'required' key, which isn't always the case. For example, the biosample profile has it, whereas the file profile has it in the anyOf subschema. For this latter case, the method now returns the empty array as there isn't at present any attempt to figure out what is conditionally required. This would have affected attempts to remove a property from such profiles since the behavior is to first prevent the user from removing required properties by popping those out of the payload.
1. Fixed bug reported by Jennifer Jou where the `profile.Profile._set_profile_id()` method didn't properly singularize the profile ID in all cases. Fixed this by using the inflection module's singularize function.
2. Fixed bug where `eu_register.py`'s `typecast()` function didn't check for booleans to typecaset to. That meant that the registration script didn't always handle boolean fields properly. Thanks again to jjou for reporting.

New in Master:

10/23/2018
1. Added script [``eu_get_accessions.py``](https://encode-utils.readthedocs.io/en/latest/scripts/eu_get_accessions.html).
Given an input list of record aliases, retrieves the DCC accession for each.
2. Added script [``eu_create_gcp_url_list.py``](https://encode-utils.readthedocs.io/en/latest/scripts/eu_create_gcp_url_list.html).

Updates in Master:

1. Documentation in transer_to_gcp.py has been updated.
1. Renamed ``utils.clean_alias_name()`` to [``utils.clean_aliases``](https://encode-utils.readthedocs.io/en/latest/utils.html#encode_utils.utils.clean_aliases). This function now takes a list of aliases and either removes or replaces non-permitted characters, such as "/" and "". This function is called in the pre-submit hook [``before_submit_alias``](https://encode-utils.readthedocs.io/en/latest/connection.html#encode_utils.connection.Connection.before_submit_alias).

New:
Added script ``eu_generate_upload_creds.py``.
Added support for creating a URL list file in ``transfer_to_gcpy.py`` and at ``connection.Connection.gcp_transfer_urllist``. This file can be used in GCP to copy files from ENCODE AWS buckets to GCP.

Updates:
Added option in ``transfer_to_gcp.py`` when transferring files to GCP to allow overwrites.
Fixed bug in ``connection.Connection.gcp_transfer`` method so that it correctly finds the s3 object path.

**Known Bugs in this release**
- The s3 to gcp transfer mechanism won't work since the s3 bucket name hard-coded to the test value of pulsar-encode-assets. This is fixed in master.

**Updates in this release**
- Added missing reference to dependency jsonschema in the setup.py file. No need to upgrade to this version if you already have release [2.3.0](https://github.com/StanfordBioinformatics/encode_utils/releases/2.3.0) and jsonschema in your Python packages list. Thanks yunhailuo for adding the dependency fix.

New:
1. Added support to copy ENCODE files in AWS S3 to GCP; see the [RTD documentation](https://encode-utils.readthedocs.io/en/latest/examples_gcp.html).
1. Added script ``eu_search_results_json.py``, which accepts an ENCODE Portal search URL and saves the results to a JSON file.

Updates:
1. ``eu_register.py`` will not accept JSON in addition to tsv, thanks to the contribution from yunhailuo.

Bug fixes:
1. See notes in [2.0.0-pre](https://github.com/StanfordBioinformatics/encode_utils/releases/tag/2.0.0-pre) pre-release.

1. Updated module transfer_to_gcp.py such that is offers a class-based API rather than functions.
1. Fixed bug that broke docs on ReadTheDocs.

New Features/Updates

1. Renamed the new module named cloud_transfers.py that appeared in the last pre-release to transfer_to_gcp.py
1. Added more functions to transfer_to_gcp.py to support querying transfers.

Bug Fixes
1. In the copy_files_to_gcp() function, made the end date of the transfer equal to the start date so that the transfer is a one-off (otherwise it becomes a repeated job that runs each day until the end date.