[FEAT] Add Mistral embedding adapter (mistral-embed)

[ahmedk20]

Mistral was a first-class LLM adapter but had no embedding adapter, despite shipping a real, popular embedding model (mistral-embed). This adds a native Mistral embedding adapter that routes through LiteLLM's mistral/ provider, so cost calculation works out of the box.

• MistralEmbeddingParameters in base1.py (idempotent mistral/ prefix, required-model validation)
• MistralEmbeddingAdapter + UI JSON schema, reusing the existing logo
• Registered in embedding1/init.py
• 28 unit tests mirroring the Gemini embedding coverage

Closes #2157

What

• Adds a native Mistral embedding adapter (mistral-embed) to unstract/sdk1. Mistral was already a first-class LLM adapter but had no embedding counterpart.

• MistralEmbeddingParameters in base1.py (idempotent mistral/ prefix, required-model validation)

• MistralEmbeddingAdapter + UI JSON schema (embedding1/static/mistral.json), reusing the existing Mistral logo

• Registered in embedding1/__init__.py

• 28 unit tests mirroring the Gemini embedding coverage

Why

• Users who standardize on Mistral for generation had to switch providers just for embeddings — a second API key, second billing relationship, and mixed providers inside one RAG pipeline.
• The gap was a plain asymmetry: 14 LLM adapters include mistral, but the 9 embedding adapters did not, despite mistral-embed being a real, popular model.
• Routing through LiteLLM's native mistral/ provider means cost calculation works out of the box (litellm_provider: "mistral" matches get_provider()).

How

• New MistralEmbeddingParameters(BaseEmbeddingParameters) validates the required model and prepends the mistral/ prefix idempotently (won't double-prefix an already-prefixed model).
• MistralEmbeddingAdapter follows the existing embedding-adapter contract (get_id/get_provider/get_adapter_type/etc.), returns AdapterTypes.EMBEDDING, and is auto-discovered by register_adapters().
• Verified against LiteLLM pricing data: mistral/mistral-embed → litellm_provider: "mistral", mode: "embedding", 8192 max tokens.

Can this PR break any existing features. If yes, please list possible items. If no, please explain why. (PS: Admins do not merge the PR without this section filled)

• No. The change is purely additive: one new adapter file, one new JSON schema, one new parameter class, and one new import line. No existing adapter, schema, or shared code path is modified. Adapters are auto-registered by directory scan, so nothing else needs to change. The new provider name (mistral) already exists as an LLM provider, so cost-calculation behavior for existing adapters is untouched.

Database Migrations

• None.

Env Config

• None. Users supply their Mistral API key through the adapter UI like any other embedding adapter.

Relevant Docs

• LiteLLM Mistral provider: https://docs.litellm.ai/docs/providers/mistral
• Model: mistral-embed (auto-prefixed to mistral/mistral-embed)

Related Issues or PRs

• Closes [FEAT] Add Mistral embedding adapter (mistral-embed) #2157

Dependencies Versions

• No new dependencies. Uses the already-vendored LiteLLM mistral/ provider.

Notes on Testing

• Added tests/test_mistral_embedding.py — 28 tests covering registration, ID format, provider name, JSON schema (required fields, defaults, password format), model-prefix idempotency, input-mutation safety, and validation errors (empty/whitespace/None/missing model, missing API key).
• Full run: 52 passed (28 new + 24 Gemini embedding as a regression check).
• Lint clean with the repo's pinned ruff 0.3.4: ruff check → all passed; ruff format --check → all files formatted.

Screenshots

N/A — no custom UI; the adapter renders through the standard auto-generated config form.

Checklist

I have read and understood the Contribution Guidelines.

[CLAassistant]

All committers have signed the CLA.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: db69058a-c40f-401c-82f3-1317339f2ae4

📥 Commits

Reviewing files that changed from the base of the PR and between cb11a6f and 16dc46f.

📒 Files selected for processing (5)

• unstract/sdk1/src/unstract/sdk1/adapters/base1.py
• unstract/sdk1/src/unstract/sdk1/adapters/embedding1/__init__.py
• unstract/sdk1/src/unstract/sdk1/adapters/embedding1/mistral.py
• unstract/sdk1/src/unstract/sdk1/adapters/embedding1/static/mistral.json
• unstract/sdk1/tests/test_mistral_embedding.py

---

Summary by CodeRabbit

• New Features

  • Added support for a new Mistral embedding adapter, including its configuration, identity, and provider details.
  • Introduced a new setup schema so it can be configured with an API key, model, and timeout settings.

• Bug Fixes

  • Improved model handling by normalizing Mistral model names and validating required fields more strictly.
  • Added coverage to ensure adapter settings are parsed consistently and invalid input is rejected.

Walkthrough

Adds a new Mistral embedding adapter to the Unstract SDK, including a MistralEmbeddingParameters class with model validation, a MistralEmbeddingAdapter class with registry metadata, package exports, a UI JSON schema, and a corresponding test suite.

Changes

Mistral Embedding Adapter

Layer / File(s)	Summary
Parameter validation unstract/sdk1/src/unstract/sdk1/adapters/base1.py	Adds MistralEmbeddingParameters with api_key and optional embed_batch_size fields, validate() returning normalized model_dump(), and validate_model() enforcing non-empty model names prefixed with mistral/.
Adapter class and registration unstract/sdk1/src/unstract/sdk1/adapters/embedding1/mistral.py, unstract/sdk1/src/unstract/sdk1/adapters/embedding1/__init__.py	Implements MistralEmbeddingAdapter with get_id, get_metadata, get_name, get_description, get_provider, get_icon, and get_adapter_type (EMBEDDING), and registers/exports it via the package __init__.
JSON configuration schema unstract/sdk1/src/unstract/sdk1/adapters/embedding1/static/mistral.json	Adds a schema defining required fields (adapter_name, api_key, model) with defaults for model, password-formatted api_key, and a constrained numeric timeout.
Test coverage unstract/sdk1/tests/test_mistral_embedding.py	Adds tests validating adapter registration, identity/provider values, schema correctness, model/API key validation behavior, and metadata output.

Estimated code review effort: 2 (Simple) | ~15 minutes

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly matches the main change: adding a Mistral embedding adapter for mistral-embed.
Description check	✅ Passed	The description includes the required sections and covers what, why, how, risks, testing, and related issues.
Linked Issues check	✅ Passed	The changes implement the requested Mistral embedding adapter, schema, provider routing, and parameter class from #2157.
Out of Scope Changes check	✅ Passed	The patch appears additive and focused on the requested adapter, schema, exports, and tests with no unrelated changes.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[greptile-apps]

Greptile Summary

This PR adds a native Mistral embedding adapter (mistral-embed) to unstract/sdk1, closing a gap where Mistral was supported as an LLM provider but lacked an embedding counterpart. The implementation faithfully mirrors the existing GeminiEmbeddingAdapter pattern and routes through LiteLLM's mistral/ provider for out-of-the-box cost calculation.

• MistralEmbeddingParameters added to base1.py with idempotent mistral/ prefix logic, empty/whitespace/None model validation, and a shallow-copy-safe validate() method consistent with GeminiEmbeddingParameters.
• MistralEmbeddingAdapter registered in embedding1/__init__.py, with a matching static JSON schema that mirrors gemini.json (same required fields, same timeout default of 240 s, password-format API key).
• 28 unit tests added covering registration, ID format, schema structure, model-prefix idempotency, mutation safety, and all validation-error paths.

Confidence Score: 5/5

Purely additive change; no existing adapter, schema, or shared code path is modified.

The new adapter faithfully mirrors the Gemini embedding adapter in every structural detail — parameter class, validate/validate_model logic, static JSON schema, registration, and test coverage. The unique adapter ID does not collide with the existing Mistral LLM adapter. The shallow-copy pattern in validate() correctly prevents input mutation, and all edge cases (None, whitespace, missing key, double-prefix) are explicitly tested.

No files require special attention.

Important Files Changed

Filename	Overview
unstract/sdk1/src/unstract/sdk1/adapters/base1.py	Adds MistralEmbeddingParameters class mirroring GeminiEmbeddingParameters; validate() correctly shallow-copies input before mutation and validate_model() handles all edge cases (None, whitespace, missing key, double-prefix). No issues.
unstract/sdk1/src/unstract/sdk1/adapters/embedding1/mistral.py	New adapter class inheriting MistralEmbeddingParameters and BaseAdapter; all required abstract methods are implemented and consistent with the Gemini counterpart. ID UUID is unique from the existing Mistral LLM adapter.
unstract/sdk1/src/unstract/sdk1/adapters/embedding1/static/mistral.json	JSON schema mirrors gemini.json exactly (same required fields, timeout default of 240 s, password format for api_key). embed_batch_size intentionally omitted from UI exposure; consistent with Gemini schema.
unstract/sdk1/src/unstract/sdk1/adapters/embedding1/init.py	Clean import and all registration of MistralEmbeddingAdapter; adapter is also auto-discovered via register_adapters directory scan.
unstract/sdk1/tests/test_mistral_embedding.py	28 well-structured tests covering registration, schema structure, all validation paths, and mutation safety. Mirrors the Gemini test suite as intended.

Class Diagram

%%{init: {'theme': 'neutral'}}%%
classDiagram
    class BaseModel {
        <<pydantic>>
    }
    class BaseEmbeddingParameters {
        <<abstract>>
        +str model
        +float|int|None timeout = 600
        +int|None max_retries = None
        +validate(adapter_metadata)* dict
        +validate_model(adapter_metadata)* str
    }
    class MistralEmbeddingParameters {
        +str api_key
        +int|None embed_batch_size = None
        +validate(adapter_metadata) dict
        +validate_model(adapter_metadata) str
    }
    class BaseAdapter {
        <<abstract>>
        +get_id()* str
        +get_name()* str
        +get_description()* str
        +get_provider()* str
        +get_icon()* str
        +get_adapter_type()* AdapterTypes
        +get_json_schema() str
        +get_metadata()* dict
    }
    class MistralEmbeddingAdapter {
        +get_id() str
        +get_name() str
        +get_description() str
        +get_provider() str
        +get_icon() str
        +get_adapter_type() AdapterTypes
        +get_metadata() dict
    }
    class GeminiEmbeddingAdapter {
        +get_id() str
        +get_provider() str
        +get_adapter_type() AdapterTypes
    }
    class GeminiEmbeddingParameters {
        +str api_key
        +int|None embed_batch_size = None
        +validate(adapter_metadata) dict
        +validate_model(adapter_metadata) str
    }

    BaseModel <|-- BaseEmbeddingParameters
    BaseEmbeddingParameters <|-- MistralEmbeddingParameters
    BaseEmbeddingParameters <|-- GeminiEmbeddingParameters
    MistralEmbeddingParameters <|-- MistralEmbeddingAdapter
    BaseAdapter <|-- MistralEmbeddingAdapter
    GeminiEmbeddingParameters <|-- GeminiEmbeddingAdapter
    BaseAdapter <|-- GeminiEmbeddingAdapter

Loading

%%{init: {'theme': 'base', 'themeVariables': {"darkMode": true, "background": "#0d1117", "primaryColor": "#21262d", "primaryTextColor": "#e6edf3", "primaryBorderColor": "#8b949e", "lineColor": "#8b949e", "textColor": "#e6edf3", "edgeLabelBackground": "#161b22", "actorBkg": "#21262d", "actorBorder": "#8b949e", "actorTextColor": "#e6edf3", "actorLineColor": "#8b949e", "signalColor": "#8b949e", "signalTextColor": "#e6edf3", "noteBkgColor": "#373320", "noteBorderColor": "#d4a72c", "noteTextColor": "#f0e6c0", "labelBoxBkgColor": "#21262d", "labelBoxBorderColor": "#8b949e", "labelTextColor": "#e6edf3", "loopTextColor": "#e6edf3", "activationBkgColor": "#30363d", "activationBorderColor": "#8b949e"}}}%%
classDiagram
    class BaseModel {
        <<pydantic>>
    }
    class BaseEmbeddingParameters {
        <<abstract>>
        +str model
        +float|int|None timeout = 600
        +int|None max_retries = None
        +validate(adapter_metadata)* dict
        +validate_model(adapter_metadata)* str
    }
    class MistralEmbeddingParameters {
        +str api_key
        +int|None embed_batch_size = None
        +validate(adapter_metadata) dict
        +validate_model(adapter_metadata) str
    }
    class BaseAdapter {
        <<abstract>>
        +get_id()* str
        +get_name()* str
        +get_description()* str
        +get_provider()* str
        +get_icon()* str
        +get_adapter_type()* AdapterTypes
        +get_json_schema() str
        +get_metadata()* dict
    }
    class MistralEmbeddingAdapter {
        +get_id() str
        +get_name() str
        +get_description() str
        +get_provider() str
        +get_icon() str
        +get_adapter_type() AdapterTypes
        +get_metadata() dict
    }
    class GeminiEmbeddingAdapter {
        +get_id() str
        +get_provider() str
        +get_adapter_type() AdapterTypes
    }
    class GeminiEmbeddingParameters {
        +str api_key
        +int|None embed_batch_size = None
        +validate(adapter_metadata) dict
        +validate_model(adapter_metadata) str
    }

    BaseModel <|-- BaseEmbeddingParameters
    BaseEmbeddingParameters <|-- MistralEmbeddingParameters
    BaseEmbeddingParameters <|-- GeminiEmbeddingParameters
    MistralEmbeddingParameters <|-- MistralEmbeddingAdapter
    BaseAdapter <|-- MistralEmbeddingAdapter
    GeminiEmbeddingParameters <|-- GeminiEmbeddingAdapter
    BaseAdapter <|-- GeminiEmbeddingAdapter

Loading

_{Reviews (1): Last reviewed commit: "[FEAT] Add Mistral embedding adapter (mi..." | Re-trigger Greptile}

[ahmedk20]

@chandrasekharan-zipstack