Add Anthropic's Claude Opus 4.7 model support

[PeterDaveHello]

Reference:

• https://www.anthropic.com/news/claude-opus-4-7
• https://platform.claude.com/docs/en/about-claude/models/overview

[qodo-free-for-open-source-projects]

Review Summary by Qodo

Add Claude Opus 4.7 model support across providers

✨ Enhancement

Walkthroughs

Description

• Add Claude Opus 4.7 model support across multiple providers
• Set context length to 1,000,000 tokens for new model
• Support Anthropic, Vertex AI, and Bedrock provider variants
• Add comprehensive unit tests for all model identifier formats

Diagram

flowchart LR
  A["Claude Opus 4.7<br/>Model Addition"] --> B["Anthropic Provider"]
  A --> C["Vertex AI Provider"]
  A --> D["Bedrock Provider"]
  B --> E["1M Token Context"]
  C --> E
  D --> E
  A --> F["Unit Tests"]
  F --> G["6 Model Variants"]

Loading

File Changes

1. pr_agent/algo/__init__.py ✨ Enhancement +6/-0

Register Claude Opus 4.7 model identifiers with context lengths

• Added vertex_ai/claude-opus-4-7 with 1,000,000 token context length
• Added anthropic/claude-opus-4-7 with 1,000,000 token context length
• Added claude-opus-4-7 with 1,000,000 token context length
• Added bedrock/anthropic.claude-opus-4-7 with 1,000,000 token context length
• Added bedrock/global.anthropic.claude-opus-4-7 and bedrock/us.anthropic.claude-opus-4-7
 variants with 1,000,000 token context length

pr_agent/algo/init.py

---

2. tests/unittest/test_get_max_tokens.py 🧪 Tests +23/-0

Add unit tests for Claude Opus 4.7 max tokens

• Added parametrized test test_claude_opus_4_7_model_max_tokens covering 6 model identifier
 formats
• Tests verify all Claude Opus 4.7 variants return 1,000,000 max tokens
• Covers Anthropic, Vertex AI, and Bedrock provider formats

tests/unittest/test_get_max_tokens.py

---

[qodo-free-for-open-source-projects]

Code Review by Qodo

🐞 Bugs (1) 📘 Rule violations (1)

1. Bedrock Opus ID mismatch 🐞 Bug ≡ Correctness

Description

The new Bedrock Claude Opus 4.7 MAX_TOKENS keys omit the -v1:0 suffix, while existing Bedrock
Claude Opus entries and the repo’s Bedrock configuration docs use -v1:0. Since get_max_tokens()
only does exact dictionary lookup, users configuring bedrock/...-v1:0 will hit an exception and
PR-Agent will fail for that model string.

Code

pr_agent/algo/init.py[156]

+    'bedrock/anthropic.claude-opus-4-7': 1000000,

Evidence

MAX_TOKENS contains Bedrock Claude Opus entries with -v1:0 for prior Opus versions, but the
newly-added Opus 4.7 Bedrock entries do not. get_max_tokens() requires an exact match in
MAX_TOKENS and raises if the key is missing, and the documentation examples for Bedrock models
include -v1:0, making a mismatch likely in real configuration.

pr_agent/algo/init.py[151-174]
pr_agent/algo/utils.py[991-1012]
docs/docs/usage-guide/changing_a_model.md[231-244]
tests/unittest/test_get_max_tokens.py[149-181]

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

### Issue description
Bedrock Claude Opus 4.7 was added to `MAX_TOKENS` without the `-v1:0` suffix, but existing Bedrock Claude Opus IDs in `MAX_TOKENS` (and the repo’s Bedrock docs/examples) use `-v1:0`. Because `get_max_tokens()` requires an exact match, using the `...-v1:0` model string will raise and break execution.

### Issue Context
- Prior Bedrock Opus entries are suffixed with `-v1:0`.
- The new Opus 4.7 Bedrock entries are not.
- Tests currently validate only the unsuffixed Opus 4.7 Bedrock strings, so they won’t catch the `-v1:0` variant failing.

### Fix Focus Areas
- pr_agent/algo/__init__.py[151-174]
- tests/unittest/test_get_max_tokens.py[149-181]

### What to change
1. Add `MAX_TOKENS` entries for the expected Bedrock Opus 4.7 IDs that include `-v1:0` (e.g., `bedrock/anthropic.claude-opus-4-7-v1:0`, `bedrock/global.anthropic.claude-opus-4-7-v1:0`, `bedrock/us.anthropic.claude-opus-4-7-v1:0`).
2. Update/extend the unit test to include these `-v1:0` variants (and keep the unsuffixed ones too if you intend to support both spellings).

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

---

2. Single quotes in fake_settings 📘 Rule violation ⚙ Maintainability

Description

The newly added test builds fake_settings using single-quoted strings, diverging from the
project’s stated preference for double quotes and the surrounding test style. This introduces
inconsistent formatting and may trigger lint/style churn in future edits.

Code

tests/unittest/test_get_max_tokens.py[R161-166]

+        fake_settings = type('', (), {
+            'config': type('', (), {
+                'custom_model_max_tokens': 0,
+                'max_model_tokens': 0
+            })()
+        })()

Evidence

PR Compliance ID 8 and 21 require adhering to repository formatting conventions (including
preferring double quotes) and keeping code compliant with lint/format tooling. The added lines use
single quotes for strings/keys in the new test setup, unlike nearby sections using double quotes.

AGENTS.md
tests/unittest/test_get_max_tokens.py[161-166]
Best Practice: Learned patterns

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

## Issue description
The new test uses single quotes for string literals and dict keys, which conflicts with the project’s preference for double quotes and increases formatting inconsistency within the file.

## Issue Context
This was introduced in the newly added `test_claude_opus_4_7_model_max_tokens` setup for `fake_settings`.

## Fix Focus Areas
- tests/unittest/test_get_max_tokens.py[161-166]

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

---

[qodo-free-for-open-source-projects]

1. Bedrock opus id mismatch 🐞 Bug ≡ Correctness

The new Bedrock Claude Opus 4.7 MAX_TOKENS keys omit the -v1:0 suffix, while existing Bedrock
Claude Opus entries and the repo’s Bedrock configuration docs use -v1:0. Since get_max_tokens()
only does exact dictionary lookup, users configuring bedrock/...-v1:0 will hit an exception and
PR-Agent will fail for that model string.

Agent Prompt

### Issue description
Bedrock Claude Opus 4.7 was added to `MAX_TOKENS` without the `-v1:0` suffix, but existing Bedrock Claude Opus IDs in `MAX_TOKENS` (and the repo’s Bedrock docs/examples) use `-v1:0`. Because `get_max_tokens()` requires an exact match, using the `...-v1:0` model string will raise and break execution.

### Issue Context
- Prior Bedrock Opus entries are suffixed with `-v1:0`.
- The new Opus 4.7 Bedrock entries are not.
- Tests currently validate only the unsuffixed Opus 4.7 Bedrock strings, so they won’t catch the `-v1:0` variant failing.

### Fix Focus Areas
- pr_agent/algo/__init__.py[151-174]
- tests/unittest/test_get_max_tokens.py[149-181]

### What to change
1. Add `MAX_TOKENS` entries for the expected Bedrock Opus 4.7 IDs that include `-v1:0` (e.g., `bedrock/anthropic.claude-opus-4-7-v1:0`, `bedrock/global.anthropic.claude-opus-4-7-v1:0`, `bedrock/us.anthropic.claude-opus-4-7-v1:0`).
2. Update/extend the unit test to include these `-v1:0` variants (and keep the unsuffixed ones too if you intend to support both spellings).

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools