Cartography

- Add Okta intel module.
- Map Okta groups to AWS roles and add Okta application reply URIs.

- Change supported version of Neo4j from 3.2 to 3.5.
- Change AWS EC2 intel module to consume query results during ingestion.
- Remove using of old Neo4j driver API.
- Include projectnumber when loading GCP projects.
- Add interesting fields to RDS databases.
- Update incorrect Neo4j index - remove :Principal and add :AWSPrincipal.
- Fix bug in GSuite intel module which could cause partial data ingestion.
- Fix syntax error in Travis file.
- Ingest Google Cloud buckets.

- Added GSuite intel module.
- Fixed CRXcavator intel module to ingest all extension data.
- Fixed incorrect linking of AWS users to groups.
- Fixed incorrect 'lastupdated' field on AWS VPC peering relationships.

- Added new lines to end of test files for Drift Detection tests are no longer inconsistent
- Drift Detection: refactored reporting to use lists instead of dicts to maintain result order
- Move GCP unit tests to tests/unit/cartography/intel
- Add support for AWS EC2 key pairs
- Fixed issue 123 - logic bug in import crxcavator extensions
- Sorted collect values in boltstatementrecords to enforce consistency with each database pull

- Added pre-commit linting
- Drift detection: allow `add_shortcuts()` to have shortcuts as arguments
- Fixed 49 - Warn and return None when a NoSuchBucket error is caught from `get_bucket_acl()`
- Added CRXcavator intel module to ingest Chrome extension data
- Fixed 104 - Drift Detection CLI failed occasionally
- Fixed 99 - improved how AWS ELBs are determined to be internet exposed
- IAM: fixed UnboundLocalError and added support for AWS IAM Federated Principals
- Added Drift Detection feature to track graph changes over time

- GCP instances that are exposed to the internet via firewall rules are now marked with `exposed_internet: True`