feat: add native Google/Gemini Embedding 2 support with Parts API

docs/en/guides/01-configuration.md

@@ -115,7 +115,7 @@ Embedding model configuration for vector search, supporting dense, sparse, and h
 | Parameter | Type | Description |
 |-----------|------|-------------|
 | `max_concurrent` | int | Maximum concurrent embedding requests (`embedding.max_concurrent`, default: `10`) |
-| `provider` | str | `"volcengine"`, `"openai"`, `"vikingdb"`, `"jina"`, or `"voyage"` |
+| `provider` | str | `"volcengine"`, `"openai"`, `"vikingdb"`, `"jina"`, `"voyage"`, or `"google"` |
 | `api_key` | str | API key |
 | `model` | str | Model name |
 | `dimension` | int | Vector dimension. For Voyage, this maps to `output_dimension` |
@@ -124,10 +124,11 @@ Embedding model configuration for vector search, supporting dense, sparse, and h
 
 **Available Models**
 
-| Model | Dimension | Input Type | Notes |
-|-------|-----------|------------|-------|
-| `doubao-embedding-vision-250615` | 1024 | multimodal | Recommended |
-| `doubao-embedding-250615` | 1024 | text | Text only |
+| Provider | Model | Dimension | Input Type | Notes |
+|----------|-------|-----------|------------|-------|
+| `volcengine` | `doubao-embedding-vision-250615` | 1024 | multimodal | Recommended |
+| `volcengine` | `doubao-embedding-250615` | 1024 | text | Text only |
+| `google` | `gemini-embedding-2-preview` | 3072 | text | Google Gemini Embedding 2 with MRL |
 
 With `input: "multimodal"`, OpenViking can embed text, images (PNG, JPG, etc.), and mixed content.
 
@@ -137,6 +138,7 @@ With `input: "multimodal"`, OpenViking can embed text, images (PNG, JPG, etc.),
 - `vikingdb`: VikingDB Embedding API
 - `jina`: Jina AI Embedding API
 - `voyage`: Voyage AI Embedding API
+- `google`: Google/Gemini AI Embedding API
 
 **vikingdb provider example:**
 
@@ -192,6 +194,29 @@ Get your API key at https://jina.ai
 }
 ```
 
+**google provider example:**
+
+```json
+{
+  "embedding": {
+    "dense": {
+      "provider": "google",
+      "api_key": "your-google-api-key",
+      "model": "gemini-embedding-2-preview",
+      "dimension": 1024,
+      "query_param": "RETRIEVAL_QUERY",
+      "document_param": "RETRIEVAL_DOCUMENT"
+    }
+  }
+}
+```
+
+For Google/Gemini embeddings:
+- `query_param` and `document_param` support task-specific embeddings
+- Valid task types: `RETRIEVAL_QUERY`, `RETRIEVAL_DOCUMENT`, `SEMANTIC_SIMILARITY`, `CLASSIFICATION`, `CLUSTERING`
+- Enhanced format: `"task_type=RETRIEVAL_QUERY,output_dimensionality=1024"`
+- Get your API key at https://aistudio.google.com/app/apikey
+
 Supported Voyage text embedding models include:
 - `voyage-4-lite`
 - `voyage-4`

docs/zh/guides/01-configuration.md

@@ -121,7 +121,7 @@ OpenViking 使用 JSON 配置文件（`ov.conf`）进行设置。配置文件支
 | 参数 | 类型 | 说明 |
 |------|------|------|
 | `max_concurrent` | int | 最大并发 Embedding 请求数（`embedding.max_concurrent`，默认：`10`） |
-| `provider` | str | `"volcengine"`、`"openai"`、`"vikingdb"` 或 `"jina"` |
+| `provider` | str | `"volcengine"`、`"openai"`、`"vikingdb"`、`"jina"`、`"voyage"` 或 `"google"` |
 | `api_key` | str | API Key |
 | `model` | str | 模型名称 |
 | `dimension` | int | 向量维度 |

openviking/models/embedder/__init__.py

@@ -13,6 +13,7 @@
 - Volcengine: Dense, Sparse, Hybrid
 - Jina AI: Dense only
 - Voyage AI: Dense only
+- Google/Gemini: Dense only
 """
 
 from openviking.models.embedder.base import (
@@ -23,9 +24,9 @@
     HybridEmbedderBase,
     SparseEmbedderBase,
 )
+from openviking.models.embedder.google_embedders import GoogleDenseEmbedder
 from openviking.models.embedder.jina_embedders import JinaDenseEmbedder
 from openviking.models.embedder.openai_embedders import OpenAIDenseEmbedder
-from openviking.models.embedder.voyage_embedders import VoyageDenseEmbedder
 from openviking.models.embedder.vikingdb_embedders import (
     VikingDBDenseEmbedder,
     VikingDBHybridEmbedder,
@@ -36,6 +37,7 @@
     VolcengineHybridEmbedder,
     VolcengineSparseEmbedder,
 )
+from openviking.models.embedder.voyage_embedders import VoyageDenseEmbedder
 
 __all__ = [
     # Base classes
@@ -45,6 +47,8 @@
     "SparseEmbedderBase",
     "HybridEmbedderBase",
     "CompositeHybridEmbedder",
+    # Google/Gemini implementations
+    "GoogleDenseEmbedder",
     # Jina AI implementations
     "JinaDenseEmbedder",
     # OpenAI implementations