[Feature Request] display more meaningful name as summary

[m1n9o]

Hi team, currently the markdown table generated by tf-summary looks like.

CHANGE	NAME
update	module.my_team.module.services["training"].pagerduty_service.this
	module.my_team.module.services["training-incident-triage"].pagerduty_service.this

However, the name module.my_team.module.services["training"].pagerduty_service.this is not a good indicator to tell me which resource. Shall we consider to refactor the output a little bit? Like

CHANGE	TYPE	NAME
update	service	Training Service

[dineshba]

Hi @m1n9o,

As far as I know, we cannot easily do that.

1. Users can name their module, resources as they want to
2. Users can have nested modules
3. Users can have the same resource with same name cross multiple modules
4. We have lot of terraform providers, each resource in each provider has some meaning

So, I feel it is not easy. Let me know if you have any ideas to achieve this. Thanks

[raghavanrrs]

use this python script meanwhile

import json
import argparse
from prettytable import PrettyTable

def parse_json(file):
    with open(file) as json_file:
        data = json.load(json_file)
    table = PrettyTable()
    table.field_names = ["CHANGE", "TYPE", "RESOURCE", "NAME"]
    for resource in data.get("resource_changes", {}):
        change = ", ".join(resource.get("change", {}).get("actions", []))
        resource_type = resource.get("type")
        resource_name = resource.get("name")
        resource_addr = resource.get("address")
        if change not in  ["no-op", "read"]:
          table.add_row([change, resource_type, resource_addr, resource_name])
    print(table)
def main():
    parser = argparse.ArgumentParser(description="Parses a Terraform JSON file.")
    parser.add_argument("--file", help="The file path to parse.")
    args = parser.parse_args()
    parse_json(args.file)
if __name__ == "__main__":
    main()

pyton tf_summary.py --file tfplan.json

output as follows:

+----------------+---------------+-------------------------+-----------+
|     CHANGE     |      TYPE     |         RESOURCE        |    NAME   |
+----------------+---------------+-------------------------+-----------+
| delete, create | null_resource | null_resource.build_jar | build_jar |
+----------------+---------------+-------------------------+-----------+

[pataquets]

Just wondering... Are HTML, Markdown, etc. formats "templatable" in any way, @dineshba?
I mean, "human readable" output formats might benefit from it. Obviously, this would exclude "machine readable" formats, since they can be processed with external tools, like jq and the likes.

Furthermore, I've not dug in the code but, if "human readable" output formats are just parsings/formattings of some "machine" format (e.g. JSON), opening the transformation pipeline to users for templating would bring powerful customization capabilities.