Trailscraper

(skipeed because of continuing release-script issues)

(same as 0.6.1, just fixing inconsistent release)

Added

* Support for Python 3.7 and 3.8
* Support for org-level trails (101)

Fixed

* `trailscraper guess` was not working when installed through homebrew or pip (110)

Removed

* **Removed official support for Python 2.7 and 3.4**. TrailScraper might still run but we no longer actively test for it

Added

* New command `guess` to extend existing policy by guessing matching actions 22

Fixed

* Fixed parsing events that contain resources without an ARN (e.g. `s3:ListObjects`) 51

**Breaking CLI changes**: split up `generate-policy` into `select` and `generate` (38)

Added

* New command `select` to print all CloudTrail records matching a filter to stdout
* New command `generate` to take CloudTrail records from stdin and generate a policy for it

Changed

* New command `select` defaults to not filtering at all whereas `generate-policy` filtered for recent events by default.
Changed to make filtering more explicit and predictable instead of surprising users who wonder why their events don't show up

Removed

* Removed command `generate-policy`, replaced with `select` and `generate`. Use pipes to produce the same behavior:
bash
$ trailscraper select | trailscraper generate

Fixed

* Made trailscraper timezone-aware. Until now, trailscraper implicitly treated everything as UTC, meaning relative timestamps (e.g. `now`, `two hours ago`) didn't work properly when filtering logfiles to download or records to generate from. (39)

Added

* New command `trailscraper last-event-timestamp` to get the last known event timestamp.
* New flag `trailscraper download --wait` to wait until events for the specified timeframe are found.
Useful if you are waiting for CloudTrail to ship logs for a recent operation.