refactor codex tool-call response translation

Merge pull request #1654 from alexey-yanchenko/feature/pass-file-inputs

Pass file input from /chat/completions and /responses to codex and claude

fix: prevent duplicate function call arguments when delta events precede done

fix: handle response.function_call_arguments.done in codex→claude streaming translator

fix: handle tool call argument streaming in Codex→OpenAI translator

Pass cache usage from codex to openai chat completions

Merge pull request #1523 from xxddff/feature/removeUserField

Update internal/translator/codex/openai/responses/codex_openai-responses_request_test.go

更新internal/translator/codex/openai/responses/codex_openai-responses_request.go

Implement test for user field deletion

Delete 'user' field from raw JSON

feat(translator): support Claude thinking type adaptive

fix(translator): normalize and restrict `stop_reason`/`finish_reason` usage

refactor(translator, executor): remove redundant `bytes.Clone` calls for improved performance

fix(gemini): support snake_case thinking config fields from Python SDK

Fixed: #1403

fix(codex): convert system role to developer for codex input

docs(translator): update Codex Claude request transform docs

refactor(codex): remove codex instructions injection support

fix(translator): handle stop_reason and MAX_TOKENS for Claude responses

feat(translator): report cached token usage in Claude output

refactor(translator): remove registry model lookups from thinking config conversions

feat(codex): enable instruction toggling and update role terminology

feat(codex): enhance input transformation and remove unused `safety_identifier` field

fix(thinking): use LookupModelInfo for model data

refactor: improve thinking logic

fix(codex): only override instructions when upstream provides them

fix(codex): only override instructions in responses for OpenCode UA

feat(codex): add OpenCode instructions based on user agent

refactor(response): enhance parameter handling for Codex to Claude conversion

Merge pull request #553 from XInTheDark/fix/builtin-tools-web-search

refactor: replace `json.Marshal` and `json.Unmarshal` with `sjson` and `gjson`

fix(thinking): align budget effort mapping across translators

fix(translator): preserve built-in tools across openai<->responses

fix(thinking): map budgets to effort for level models

fix(codex): raise default reasoning effort to medium

feat(antigravity): add function ID to FunctionCall and FunctionResponse models

**refactor(translator): remove `service_tier` from Codex OpenAI request payload**

fix: use underscore suffix in short name mapping

fix(claude translator): guard tool schema properties

Fixed: #193

Fixed: #140 #133 #80

refactor(translator): remove unused log dependency and comment out debug logging

docs: add GPT-5 Codex guidelines for internal usage

refactor: streamline ConvertCodexResponseToGeminiNonStream by removing unnecessary buffer and improving response handling

Fixed: #137

Fixed #91

Fixed: #121

feat(translator): improve Claude request handling with enhanced content processing