Unrelated Resources Incorrectly Flagged as Removed in Plan

[FayizMohideen]

Bug Report: openinfraquote incorrectly identifies data resources as removed

Summary

openinfraquote is incorrectly identifying Terraform unrelated resources as being removed when processing a plan.json file. This leads to erroneous cost estimations because these resources aren't part of actual infrastructure changes. The bug seems to be in how the tool processes the plan, as it's considering unrelated blocks instead of focusing only on resource_changes with create or delete actions.
The issue specifically is that it includes prices for data block changes that can cause the final value to be incorrect.

---

Steps to Reproduce

• Unsure of the steps to reproduce but it seems to have something to do with data resource blocks.
• A plan file can be shared privately to be used to demonstrate this behaviour.
• This behaviour may be related to the depends_on block of resources outputted by oiq match

Steps to Identify this behavior

1. Generate a Terraform plan JSON file:
   terraform plan -out=tfplan && terraform show -json tfplan > plan.json
2. Run openinfraquote against the plan.json file:
   oiq match --pricesheet prices.csv plan.json
3. Filter the output to see removed resources:
   oiq match --pricesheet prices.csv plan.json | jq '.matches[] | select(.change == "remove") | .resource.address'

---

Expected Behavior

openinfraquote should only consider resources with a change_actions of create or delete from the resource_changes block. The output should not include any unrelated resources.

---

Actual Behavior

The jq command reveals that openinfraquote lists unrelated resources as being removed, which is incorrect.

---

Proposed Fix

Modify the oiq parser to only look at resource_changes[] in the plan json. This will ensure that only genuine infrastructure changes are considered for cost analysis.

[FayizMohideen]

I think a change at this line:

openinfraquote/src/oiq/oiq_tf.ml

Line 96 in fa94085

>>= fun resources -> Ok (resources @ CCList.flatten child_resources)

to filter out data resources would be sufficient to fix this issue
currently using this command in my workflow which effectively filters out data resources: ./oiq match --pricesheet prices.csv "${{ steps.find_plan.outputs.plan_path }}" | jq 'del(.matches[] | select(.resource.address | startswith("data.")))' | ./oiq price $REGION_FLAG --format=markdown > oiq-comment.md