TLX

Tools for working with AWS.

Install

This package is not distributed through pypi. Clone the repository and install it locally.

pip install -e .

CLI apps

Tools that should be in the awscli but aren't or don't function well. All have --help for details.

All CLI applications have their own detailed help menus. Currently available tools are:

function	description
`get-aws-creds`	returns temporary session credentials. Locally mock AWS runtime environments, debugging IAM etc
`dynamo-batch-write`	loads scan results into a dynamo table. Much better than `awscli` option
`dynamo-clear-table`	empties the items from a dynamodb table

$ dynamo-batch-write --help
Usage: dynamo-batch-write [OPTIONS]

  DYNAMO BATCH WRITE

  Loads the results of a scan opperation into a table.

  Details:
  Takes the output of a `scan` operation such as: `aws dynamodb scan --table-name <TableName>`
  and writes to an existing table. Similar to the `aws dynamodb batch-write-item` command except:
      - No limit to amount of items in the upload (25 with awscli)
      - Take the output of a table scan, requiring no reformatting

Options:
  -d, --dump-file FILENAME  File dumped from dynamodb with a `scan`.  [required]
  -t, --table TEXT          Table to send data to. Table must exist and key schema must match.  Use `aws dynamodb
                            describe-table --table-name <TableName>`  [required]
  -h, --help                Show this message and exit.

AWS CLI Wrappers

A collection of tools and alias' that make remembering aws cli commands easier and faster.

They are not production grade.

Standard features

They start with the name of the service so that you can hit tab and see the options you expect. E.g: code-pipeline or ecs. Some may have a short alias such as cp- for code-pipeline because this is meant to make the aws cli more intuitive. The command you want, should be guessable.
If the usage isn't obvious (e.g ecs-clusters ) then they have a --help with an example.
They produce json output. This allows standard tools like jq or jtbl to format it. The --help says what kind of output is produced. Lists of things should produce jsonlines so that the output can be grep'ed and still passed to jtbl. E.g stack-status | grep meta | jtbl

Instalation

It is a manual step to source ./tools/rcfile because you may choose to run it in a sub-shell to avoid polluting your environment. For example I add the following alias to my .bashrc.

alias awsenv='bash --rcfile ${TLXPATH}/tools/rcfile'

Otherwise source the .tools/rcfile in your own way.

examples

Find differences to be deployed by a Cope-Pipeline deployment

Find your executionId's to compare somehow. Maybe by using cp-state $p to look at different stages.

Maybe you want to look at previous executions: cp-execs $p 20 | grep -e 'Prod_Environment' | jtbl to see when Prod was last deployed.

cp-state $p to show pipeline executionId between your stages (blocked by an approval)
cp-execs $p 30 | grep "exId1\|exId2" | jtbl to get the Source hash.
git log diff1..diff2

Module Summary

Import these in a python program or shell.

function	description
`tlx.apigateway`	Reduce boilerplate when using proxy response lambdas with API Gateway
`tlx.dynamodb`	`clear_table`, batch loaders for csv bigquery and functions for nested data reliably
`tlx.util`	Extra tools such as: better boto3 `Session`, generic paginator that works on all boto3 calls + more

See Submodule docs for more examples.

tlelson / tlx

readme

TLX