Improve long coder summary rule tables

[Psuedonerd]

Motivation

• Fix Markdown rendering for the long Dolan2015 and Zhang2023 coder-summary rule tables.
• Reduce repeated boilerplate in the per-rule technical meaning column while preserving the complete rule inventories.

Description

• Removed blank lines between adjacent rule-table rows in data/summaries_coders/Dolan2015_coder.md and data/summaries_coders/Zhang2023_coder.md so GitHub/Markdown renderers treat each inventory as a contiguous table.
• Rewrote the generated technical-meaning text for the two long inventories to identify concrete operations where possible: production, degradation/sink routing, explicit cleanup, state changes, bond formation/release, and site-level contact changes.
• Kept the complete rule inventories intact: Dolan2015 remains 2,339 rule rows and Zhang2023 remains 232 logical rule rows.

Testing

• python -m json.tool data/ai_summaries.json >/tmp/ai_summaries.valid.json
• Custom table validation confirmed both long coder summaries have contiguous rule rows with no blank lines between adjacent table rows and the expected rule counts.
• Custom index validation confirmed 32 summary entries and that all referenced Markdown paths exist.
• git diff --check

---

Codex Task

[chatgpt-codex-connector]

💡 Codex Review

RuleHub/data/summaries/Blinovran_aigenerated.md

Line 5 in c021b8e

A compact Ran-system example focused on nucleotide-state and compartment-related molecular cycling.

Remove unsupported Ran nucleotide-state claims

For the Blinovran model, this summary invents nucleotide/activity-state cycling: the source declares Ran(cargo) and only gives internal ~u/~p states to the cargo molecule C(Y1,Y2,Y3), while the rules move Ran-cargo between compartments and phosphorylate cargo. Readers of the generated Markdown and matching JSON entry are therefore told the model has Ran nucleotide states and Ran state conversions that are not present in Published/Blinovran/Blinov_ran.bngl.

---

RuleHub/data/ai_summaries.json

Line 213 in c021b8e

"model_id": "lac_operon_dreisigmeyer2008",

Use metadata ids for summary lookup

For models whose BNGL filename differs from the metadata id, this index stores the BNGL basename as model_id (for example this entry should be Dreisigmeyer_2008 from Published/Dreisigmeyer2008/metadata.yaml). The generated manifests and the rest of RuleHub identify these models by metadata ids, so consumers joining data/ai_summaries.json to manifest entries by model_id will fail for this and the similar Gardner/Goldstein/Harmon/Hlavacek entries.

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".