symbolicai 1.5.0__tar.gz → 1.6.0__tar.gz

{symbolicai-1.5.0 → symbolicai-1.6.0}/AGENTS.md

@@ -53,7 +53,7 @@ CLI entrypoints (after install): `symchat`, `symsh`, `symconfig`, `symserver`.
 
 ## Configuration & Secrets
 - Config precedence: `./symai.config.json`, `{venv}/.symai/symai.config.json`, then `~/.symai/symai.config.json`.
-- Common keys: `NEUROSYMBOLIC_ENGINE_MODEL`, `NEUROSYMBOLIC_ENGINE_API_KEY`, `SYMAI_WARNINGS=0`, `SUPPORT_COMMUNITY`.
+- Common keys: `NEUROSYMBOLIC_ENGINE_MODEL`, `NEUROSYMBOLIC_ENGINE_API_KEY`, `SYMAI_WARNINGS=0`.
 - Inspect active config with `symconfig`.
 - Never commit API keys, tokens, or generated artifacts (`dist/`, caches, logs).
 

{symbolicai-1.5.0 → symbolicai-1.6.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: symbolicai
-Version: 1.5.0
+Version: 1.6.0
 Summary: A Neurosymbolic Perspective on Large Language Models
 Author-email: Marius-Constantin Dinu <marius@extensity.ai>, Leoveanu-Condrei Claudiu <leo@extensity.ai>
 License: BSD 3-Clause License
@@ -88,10 +88,6 @@ Requires-Dist: llvmlite>=0.45.1
 Requires-Dist: cerebras-cloud-sdk>=1.59.0
 Provides-Extra: bitsandbytes
 Requires-Dist: bitsandbytes>=0.43.1; extra == "bitsandbytes"
-Provides-Extra: blip2
-Requires-Dist: decord>=0.6.0; extra == "blip2"
-Requires-Dist: salesforce-lavis>=1.0.0; extra == "blip2"
-Requires-Dist: opencv-python-headless>=4.5.5.64; extra == "blip2"
 Provides-Extra: hf
 Requires-Dist: transformers>=4.45.2; extra == "hf"
 Requires-Dist: accelerate>=0.33.0; extra == "hf"
@@ -425,15 +421,12 @@ Example of a configuration file with all engines enabled:
     "VISION_ENGINE_MODEL": "openai/clip-vit-base-patch32",
     "OCR_ENGINE_API_KEY": "<APILAYER_API_KEY>",
     "SPEECH_TO_TEXT_ENGINE_MODEL": "turbo",
-    "SPEECH_TO_TEXT_API_KEY": "",
-    "SUPPORT_COMMUNITY": true
+    "SPEECH_TO_TEXT_API_KEY": ""
 }
 ```
 
 With these steps completed, you should be ready to start using SymbolicAI in your projects.
 
-> ❗️**NOTE**❗️Our framework allows you to support us train models for local usage by enabling the data collection feature. On application startup we show the terms of services and you can activate or disable this community feature. We do not share or sell your data to 3rd parties and only use the data for research purposes and to improve your user experience. To change this setting open the `symai.config.json` and turn it on/off by setting the `SUPPORT_COMMUNITY` property to `True/False` via the config file or the respective environment variable.
-
 > ❗️**NOTE**❗️By default, the user warnings are enabled. To disable them, export `SYMAI_WARNINGS=0` in your environment variables.
 
 ### Running tests

{symbolicai-1.5.0 → symbolicai-1.6.0}/README.md

@@ -282,15 +282,12 @@ Example of a configuration file with all engines enabled:
     "VISION_ENGINE_MODEL": "openai/clip-vit-base-patch32",
     "OCR_ENGINE_API_KEY": "<APILAYER_API_KEY>",
     "SPEECH_TO_TEXT_ENGINE_MODEL": "turbo",
-    "SPEECH_TO_TEXT_API_KEY": "",
-    "SUPPORT_COMMUNITY": true
+    "SPEECH_TO_TEXT_API_KEY": ""
 }
 ```
 
 With these steps completed, you should be ready to start using SymbolicAI in your projects.
 
-> ❗️**NOTE**❗️Our framework allows you to support us train models for local usage by enabling the data collection feature. On application startup we show the terms of services and you can activate or disable this community feature. We do not share or sell your data to 3rd parties and only use the data for research purposes and to improve your user experience. To change this setting open the `symai.config.json` and turn it on/off by setting the `SUPPORT_COMMUNITY` property to `True/False` via the config file or the respective environment variable.
-
 > ❗️**NOTE**❗️By default, the user warnings are enabled. To disable them, export `SYMAI_WARNINGS=0` in your environment variables.
 
 ### Running tests

{symbolicai-1.5.0 → symbolicai-1.6.0}/docs/source/ENGINES/drawing_engine.md

@@ -1,9 +1,10 @@
 # Drawing Engine
 
-We now expose two distinct high-level drawing interfaces:
+We now expose three distinct high-level drawing interfaces:
 
 1. **`gpt_image`** – a unified wrapper around OpenAI’s Images API (DALL·E 2/3 and `gpt-image-*`).
 2. **`flux`** – Black Forest Labs’ Flux text-to-image models via api.us1.bfl.ai.
+3. **`nanobanana`** – Google Gemini image generation models via `google-genai`.
 
 Both return a list of local PNG file paths.
 
@@ -153,3 +154,44 @@ Under the hood Flux uses:
 - GET  `https://api.us1.bfl.ai/v1/get_result?id={request_id}`
 
 and writes out local PNG file(s).
+
+---
+
+## 3. Google “nanobanana” (Gemini Image) Interface
+
+Use `Interface("nanobanana")` to generate images with Gemini image models (via the `google-genai` SDK).
+This interface currently supports **create-only** generation.
+
+Supported models (as of this release):
+- `gemini-2.5-flash-image`
+- `gemini-3-pro-image-preview`
+
+```python
+from symai.interfaces import Interface
+
+nanobanana = Interface("nanobanana")
+
+paths = nanobanana(
+    "a fluffy cat with a cowboy hat",
+    operation="create",                 # currently only 'create' is implemented
+    model="gemini-2.5-flash-image",
+)
+
+print(paths[0])  # → /tmp/tmpabcd.png
+```
+
+### Supported Parameters
+
+- `prompt` (str)
+- `operation` (`"create"`)
+- `model` (str, default from `SYMAI_CONFIG["DRAWING_ENGINE_MODEL"]`)
+- `response_modalities` (list[str], default `["IMAGE"]`)
+- `config` (optional): a `google.genai.types.GenerateContentConfig` instance
+- `except_remedy` (callable)
+
+### Configuration
+
+Set these keys in `symai.config.json` (or via your preferred config location):
+
+- `DRAWING_ENGINE_API_KEY`: your Gemini API key
+- `DRAWING_ENGINE_MODEL`: one of the supported Gemini image model names, e.g. `gemini-2.5-flash-image`

{symbolicai-1.5.0 → symbolicai-1.6.0}/docs/source/ENGINES/neurosymbolic_engine.md

@@ -8,6 +8,7 @@ Depending on which backend you configure (OpenAI/GPT, Claude, Gemini, Deepseek,
 * Local engines (llamacpp, HuggingFace) do *not* yet support token counting, JSON format enforcement, or vision inputs in the same way.
 * Groq engine requires a special format for the `NEUROSYMBOLIC_ENGINE_MODEL` key: `groq:model_id`. E.g., `groq:qwen/qwen3-32b`.
 * Cerebras engine requires a special format for the `NEUROSYMBOLIC_ENGINE_MODEL` key: `cerebras:model_id`. E.g., `cerebras:gpt-oss-120b`.
+* OpenRouter engine requires a special format for the `NEUROSYMBOLIC_ENGINE_MODEL` key: `openrouter:model_id`. E.g., `openrouter:moonshotai/kimi-k2.5`.
 * OpenAI Responses API engine requires the `responses:` prefix: `responses:model_id`. E.g., `responses:gpt-4.1`, `responses:o3-mini`. This uses OpenAI's newer `/v1/responses` endpoint instead of `/v1/chat/completions`.
 * Token‐truncation and streaming are handled automatically but may vary in behavior by engine.
 
@@ -209,6 +210,23 @@ print(metadata["thinking"])
 
 For OpenAI Responses API with reasoning models (e.g., `o3-mini`, `o3`, `o4-mini`, `gpt-5`, `gpt-5.1`), the thinking trace is extracted from the reasoning summary in the response output.
 
+### OpenRouter
+
+```python
+from symai import Symbol
+
+# openrouter:moonshotai/kimi-k2.5
+res, metadata = Symbol("Topic: Disneyland") \
+    .query(
+      "Write a dystopic take on the topic.",
+      return_metadata=True
+    )
+print(res)
+print(metadata.get("thinking", "No thinking trace available"))
+```
+
+OpenRouter backends provide access to multiple model providers through a unified API gateway. The engine automatically handles model routing and supports provider-specific features when available. Token counting is not implemented for OpenRouter models, similar to Groq and Cerebras engines.
+
 ---
 
 ## JSON‐Only Responses
@@ -264,7 +282,13 @@ print(Symbol(string).tokens)
 ### Tracking Usage and Estimating Costs with `MetadataTracker`
 
 For more detailed tracking of API calls, token usage, and estimating costs, you can use the `MetadataTracker` in conjunction with `RuntimeInfo`. This is particularly useful for monitoring multiple calls within a specific code block.
-> ❗️**NOTE**❗️we only track OpenAI models for now (chat and search).
+> ❗️**NOTE**❗️`MetadataTracker` collects raw per-call metadata for any `Engine`, but **token usage extraction** (i.e. `tracker.usage` → `RuntimeInfo`) is currently implemented for:
+>
+> - **OpenAI**: `GPTXChatEngine`, `GPTXReasoningEngine`, `OpenAIResponsesEngine`, `GPTXSearchEngine` (eg. `gpt-5-chat-latest`)
+> - **Claude (Anthropic)**: `ClaudeXChatEngine`, `ClaudeXReasoningEngine` (eg. `claude-sonnet-4-5`)
+> - **Gemini (Google)**: `GeminiXReasoningEngine` (e.g. `gemini-2.5-pro`, `gemini-2.5-flash`)
+>
+> For other engines, `tracker.metadata` will still contain raw outputs, but `tracker.usage` may be empty or partial.
 
 `MetadataTracker` collects metadata from engine calls made within its context. `RuntimeInfo` then processes this raw metadata to provide a summary of token counts, number of API calls, elapsed time, and an estimated cost if pricing information is provided.
 

{symbolicai-1.5.0 → symbolicai-1.6.0}/docs/source/INSTALLATION.md

@@ -144,15 +144,12 @@ Example of a configuration file with all engines enabled:
     "VISION_ENGINE_MODEL": "openai/clip-vit-base-patch32",
     "OCR_ENGINE_API_KEY": "<APILAYER_API_KEY>",
     "SPEECH_TO_TEXT_ENGINE_MODEL": "turbo",
-    "SPEECH_TO_TEXT_API_KEY": "",
-    "SUPPORT_COMMUNITY": true
+    "SPEECH_TO_TEXT_API_KEY": ""
 }
 ```
 
 With these steps completed, you should be ready to start using SymbolicAI in your projects.
 
-> ❗️**NOTE**❗️Our framework allows you to support us train models for local usage by enabling the data collection feature. On application startup we show the terms of services and you can activate or disable this community feature. We do not share or sell your data to 3rd parties and only use the data for research purposes and to improve your user experience. To change this setting open the `symai.config.json` and turn it on/off by setting the `SUPPORT_COMMUNITY` property to `True/False` via the config file or the respective environment variable.
-
 > ❗️**NOTE**❗️By default, the user warnings are enabled. To disable them, export `SYMAI_WARNINGS=0` in your environment variables.
 
 ### Running tests

{symbolicai-1.5.0 → symbolicai-1.6.0}/pyproject.toml

@@ -72,7 +72,6 @@ dependencies = [
 
 [project.optional-dependencies]
 bitsandbytes = ["bitsandbytes>=0.43.1"] # handle separately because of Apple Silicon
-blip2        = ["decord>=0.6.0", "salesforce-lavis>=1.0.0", "opencv-python-headless>=4.5.5.64"]
 hf           = ["transformers>=4.45.2", "accelerate>=0.33.0", "peft>=0.13.1", "datasets>=3.0.1", "trl>=0.11.3"]
 scrape       = ["beautifulsoup4>=4.12.3", "trafilatura>=2.0.0", "pdfminer.six", "playwright>=1.45.0", "parallel-web>=0.3.3"]
 llama_cpp    = ["llama-cpp-python[server]>=0.3.7"] # handle separately since this dependency may not compile and require special maintenance

{symbolicai-1.5.0 → symbolicai-1.6.0}/symai/__init__.py

@@ -11,8 +11,6 @@ from rich.table import Table
 from rich.tree import Tree
 
 from .backend import settings
-from .menu.screen import show_intro_menu
-from .misc.console import ConsoleStyle
 from .utils import UserMessage
 
 # do not remove - hides the libraries' debug messages
@@ -33,7 +31,7 @@ os.environ["TOKENIZERS_PARALLELISM"] = "false"
 # Create singleton instance
 config_manager = settings.SymAIConfig()
 
-SYMAI_VERSION = "1.5.0"
+SYMAI_VERSION = "1.6.0"
 __version__ = SYMAI_VERSION
 __root_dir__ = config_manager.config_dir
 
@@ -87,20 +85,6 @@ def _start_symai():
     # Load and manage configurations
     symai_config = config_manager.load_config("symai.config.json")
 
-    # MIGRATE THE ENVIRONMENT VARIABLES
-    # *==========================================================================================================*
-    if "COLLECTION_URI" not in symai_config:
-        updates = {
-            "COLLECTION_URI": "mongodb+srv://User:vt3epocXitd6WlQ6@extensityai.c1ajxxy.mongodb.net/?retryWrites=true&w=majority",
-            "COLLECTION_DB": "ExtensityAI",
-            "COLLECTION_STORAGE": "SymbolicAI",
-            "SUPPORT_COMMUNITY": False,
-        }
-        config_manager.migrate_config("symai.config.json", updates)
-        with ConsoleStyle("info") as console:
-            msg = "Currently you are sharing your user experience with us by uploading the data to our research server, and thereby helping us improve future models and the overall SymbolicAI experience. We thank you very much for supporting the research community! If you wish to disable the data collection option go to your .symai config situated in your home directory or set the environment variable `SUPPORT_COMMUNITY` to `False`."
-            console.print(msg)
-
     # POST-MIGRATION CHECKS
     # *==============================================================================================================*
     if "TEXT_TO_SPEECH_ENGINE_API_KEY" not in symai_config:
@@ -114,11 +98,6 @@ def _start_symai():
     symsh_config = config_manager.load_config("symsh.config.json")
     symserver_config = config_manager.load_config("symserver.config.json")
 
-    # MIGRATE THE SHELL SPLASH SCREEN CONFIGURATION
-    # *==============================================================================================================*
-    if "show-splash-screen" not in symsh_config:
-        config_manager.migrate_config("symsh.config.json", {"show-splash-screen": True})
-
     # CHECK IF THE USER HAS A NEUROSYMBOLIC API KEY
     # *==============================================================================================================*
     if not (
@@ -130,7 +109,6 @@ def _start_symai():
     ):
         # Try to fallback to the global (home) config if environment is not home
         if config_manager.config_dir != config_manager._home_config_dir:
-            show_intro_menu()
             UserMessage(
                 f"You didn't configure your environment ({config_manager.config_dir})! Falling back to the global ({config_manager._home_config_dir}) configuration if it exists."
             )
@@ -345,57 +323,29 @@ def display_config():
 
 
 def setup_wizard(_symai_config_path_):
-    show_intro_menu()
-
-    _nesy_engine_api_key = ""
-    _nesy_engine_model = ""
-    _symbolic_engine_api_key = ""
-    _symbolic_engine_model = ""
-    _embedding_engine_api_key = ""
-    _embedding_model = ""
-    _drawing_engine_api_key = ""
-    _drawing_engine_model = ""
-    _vision_engine_model = ""
-    _search_engine_api_key = ""
-    _search_engine_model = ""
-    _ocr_engine_api_key = ""
-    _speech_to_text_engine_model = ""
-    _speech_to_text_api_key = ""
-    _text_to_speech_engine_api_key = ""
-    _text_to_speech_engine_model = ""
-    _text_to_speech_engine_voice = ""
-    _indexing_engine_api_key = ""
-    _indexing_engine_environment = ""
-    _caption_engine_environment = ""
-    _support_comminity = False
-
     config_manager.save_config(
         _symai_config_path_,
         {
-            "NEUROSYMBOLIC_ENGINE_API_KEY": _nesy_engine_api_key,
-            "NEUROSYMBOLIC_ENGINE_MODEL": _nesy_engine_model,
-            "SYMBOLIC_ENGINE_API_KEY": _symbolic_engine_api_key,
-            "SYMBOLIC_ENGINE": _symbolic_engine_model,
-            "EMBEDDING_ENGINE_API_KEY": _embedding_engine_api_key,
-            "EMBEDDING_ENGINE_MODEL": _embedding_model,
-            "DRAWING_ENGINE_API_KEY": _drawing_engine_api_key,
-            "DRAWING_ENGINE_MODEL": _drawing_engine_model,
-            "VISION_ENGINE_MODEL": _vision_engine_model,
-            "SEARCH_ENGINE_API_KEY": _search_engine_api_key,
-            "SEARCH_ENGINE_MODEL": _search_engine_model,
-            "OCR_ENGINE_API_KEY": _ocr_engine_api_key,
-            "SPEECH_TO_TEXT_ENGINE_MODEL": _speech_to_text_engine_model,
-            "SPEECH_TO_TEXT_API_KEY": _speech_to_text_api_key,
-            "TEXT_TO_SPEECH_ENGINE_API_KEY": _text_to_speech_engine_api_key,
-            "TEXT_TO_SPEECH_ENGINE_MODEL": _text_to_speech_engine_model,
-            "TEXT_TO_SPEECH_ENGINE_VOICE": _text_to_speech_engine_voice,
-            "INDEXING_ENGINE_API_KEY": _indexing_engine_api_key,
-            "INDEXING_ENGINE_ENVIRONMENT": _indexing_engine_environment,
-            "CAPTION_ENGINE_MODEL": _caption_engine_environment,
-            "COLLECTION_URI": "mongodb+srv://User:vt3epocXitd6WlQ6@extensityai.c1ajxxy.mongodb.net/?retryWrites=true&w=majority",
-            "COLLECTION_DB": "ExtensityAI",
-            "COLLECTION_STORAGE": "SymbolicAI",
-            "SUPPORT_COMMUNITY": _support_comminity,
+            "NEUROSYMBOLIC_ENGINE_API_KEY": "",
+            "NEUROSYMBOLIC_ENGINE_MODEL": "",
+            "SYMBOLIC_ENGINE_API_KEY": "",
+            "SYMBOLIC_ENGINE": "",
+            "EMBEDDING_ENGINE_API_KEY": "",
+            "EMBEDDING_ENGINE_MODEL": "",
+            "DRAWING_ENGINE_API_KEY": "",
+            "DRAWING_ENGINE_MODEL": "",
+            "VISION_ENGINE_MODEL": "",
+            "SEARCH_ENGINE_API_KEY": "",
+            "SEARCH_ENGINE_MODEL": "",
+            "OCR_ENGINE_API_KEY": "",
+            "SPEECH_TO_TEXT_ENGINE_MODEL": "",
+            "SPEECH_TO_TEXT_API_KEY": "",
+            "TEXT_TO_SPEECH_ENGINE_API_KEY": "",
+            "TEXT_TO_SPEECH_ENGINE_MODEL": "",
+            "TEXT_TO_SPEECH_ENGINE_VOICE": "",
+            "INDEXING_ENGINE_API_KEY": "",
+            "INDEXING_ENGINE_ENVIRONMENT": "",
+            "CAPTION_ENGINE_MODEL": "",
         },
     )
 

{symbolicai-1.5.0 → symbolicai-1.6.0}/symai/backend/base.py

@@ -3,21 +3,11 @@ import time
 from abc import ABC, abstractmethod
 from typing import Any
 
-from ..collect import CollectionRepository, rec_serialize
 from ..utils import UserMessage
 from .settings import HOME_PATH
 
 ENGINE_UNREGISTERED = "<UNREGISTERED/>"
 
-COLLECTION_LOGGING_ENGINES = {
-    "GPTXChatEngine",
-    "GPTXCompletionEngine",
-    "SerpApiEngine",
-    "WolframAlphaEngine",
-    "SeleniumEngine",
-    "OCREngine",
-}
-
 
 class Engine(ABC):
     def __init__(self) -> None:
@@ -26,8 +16,6 @@ class Engine(ABC):
         self.logging = False
         self.log_level = logging.DEBUG
         self.time_clock = False
-        self.collection = CollectionRepository()
-        self.collection.connect()
         # create formatter
         __root_dir__ = HOME_PATH
         __root_dir__.mkdir(parents=True, exist_ok=True)
@@ -66,9 +54,6 @@ class Engine(ABC):
         if self.logging:
             self.logger.log(self.log_level, log)
 
-        if str(self) in COLLECTION_LOGGING_ENGINES:
-            self._record_collection_entry(argument, metadata, req_time)
-
         self._trigger_output_handlers(argument, res, metadata)
         return res, metadata
 
@@ -92,17 +77,6 @@ class Engine(ABC):
         if argument_handler:
             argument_handler((result, metadata))
 
-    def _record_collection_entry(self, argument: Any, metadata: dict, req_time: float) -> None:
-        self.collection.add(
-            forward={"args": rec_serialize(argument.args), "kwds": rec_serialize(argument.kwargs)},
-            engine=str(self),
-            metadata={
-                "time": req_time,
-                "data": rec_serialize(metadata),
-                "argument": rec_serialize(argument),
-            },
-        )
-
     def id(self) -> str:
         return ENGINE_UNREGISTERED
 

symbolicai-1.6.0/symai/backend/engines/drawing/engine_gemini_image.py

@@ -0,0 +1,101 @@
+import base64
+import logging
+import mimetypes
+import tempfile
+from pathlib import Path
+
+from google import genai
+from google.genai import types
+
+from ....symbol import Result
+from ....utils import UserMessage
+from ...base import Engine
+from ...settings import SYMAI_CONFIG
+
+logging.getLogger("google.genai").setLevel(logging.ERROR)
+logging.getLogger("google_genai").propagate = False
+
+
+class GeminiImageResult(Result):
+    def __init__(self, value, **kwargs):
+        super().__init__(value, **kwargs)
+        paths = []
+        for candidate in getattr(value, "candidates", []) or []:
+            content = getattr(candidate, "content", None)
+            parts = getattr(content, "parts", []) if content else []
+            for part in parts:
+                inline_data = getattr(part, "inline_data", None)
+                if inline_data is None:
+                    continue
+                mime_type = getattr(inline_data, "mime_type", None) or "image/png"
+                data = getattr(inline_data, "data", None)
+                if data is None:
+                    continue
+                if isinstance(data, str):
+                    data = base64.b64decode(data)
+                suffix = mimetypes.guess_extension(mime_type) or ".png"
+                with tempfile.NamedTemporaryFile(suffix=suffix, delete=False) as tmp_file:
+                    path = tmp_file.name
+                with Path(path).open("wb") as f:
+                    f.write(data)
+                paths.append(path)
+        if not paths:
+            UserMessage("Gemini image generation returned no images.", raise_with=ValueError)
+        self._value = paths
+
+
+class GeminiImageEngine(Engine):
+    def __init__(self, api_key: str | None = None, model: str | None = None):
+        super().__init__()
+        self.config = SYMAI_CONFIG
+        self.api_key = self.config.get("DRAWING_ENGINE_API_KEY") if api_key is None else api_key
+        self.model = self.config.get("DRAWING_ENGINE_MODEL") if model is None else model
+        self.name = self.__class__.__name__
+        self.client = genai.Client(api_key=self.api_key)
+
+    def id(self) -> str:
+        cfg_model = self.config.get("DRAWING_ENGINE_MODEL")
+        if cfg_model and cfg_model.startswith(("gemini-2.5-flash-image", "gemini-3-pro-image-preview")):
+            return "drawing"
+        return super().id()
+
+    def command(self, *args, **kwargs):
+        super().command(*args, **kwargs)
+        if "DRAWING_ENGINE_API_KEY" in kwargs:
+            self.api_key = kwargs["DRAWING_ENGINE_API_KEY"]
+            self.client = genai.Client(api_key=self.api_key)
+        if "DRAWING_ENGINE_MODEL" in kwargs:
+            self.model = kwargs["DRAWING_ENGINE_MODEL"]
+
+    def prepare(self, argument):
+        argument.prop.prepared_input = str(argument.prop.processed_input)
+
+    def forward(self, argument):
+        prompt = argument.prop.prepared_input
+        kwargs = argument.kwargs
+        model = kwargs.get("model", self.model)
+        operation = kwargs.get("operation")
+
+        if operation != "create":
+            UserMessage(f"Unknown operation: {operation}", raise_with=ValueError)
+
+        response_modalities = kwargs.get("response_modalities", ["IMAGE"])
+        config = kwargs.get("config")
+        if config is None:
+            config = types.GenerateContentConfig(response_modalities=response_modalities)
+
+        except_remedy = kwargs.get("except_remedy", None)
+        try:
+            res = self.client.models.generate_content(
+                model=model,
+                contents=prompt,
+                config=config,
+            )
+        except Exception as e:
+            if except_remedy is None:
+                raise
+            res = except_remedy(self, e, None, argument)
+
+        metadata = {}
+        result = GeminiImageResult(res)
+        return [result], metadata

{symbolicai-1.5.0 → symbolicai-1.6.0}/symai/backend/engines/embedding/engine_openai.py

@@ -20,25 +20,28 @@ class EmbeddingEngine(Engine, OpenAIMixin):
         logger = logging.getLogger("openai")
         logger.setLevel(logging.WARNING)
         self.config = SYMAI_CONFIG
+        self._api_key = api_key or self.config.get("EMBEDDING_ENGINE_API_KEY")
+        self._model = model or self.config.get("EMBEDDING_ENGINE_MODEL")
         if self.id() != "embedding":
             return  # do not initialize if not embedding; avoids conflict with llama.cpp check in EngineRepository.register_from_package
-        openai.api_key = self.config["EMBEDDING_ENGINE_API_KEY"] if api_key is None else api_key
-        self.model = self.config["EMBEDDING_ENGINE_MODEL"] if model is None else model
+        # Use openai client instance (required for openai 1.0+)
+        self._client = openai.OpenAI(api_key=self._api_key)
+        self.model = self._model
         self.max_tokens = self.api_max_context_tokens()
         self.embedding_dim = self.api_embedding_dims()
         self.name = self.__class__.__name__
 
     def id(self) -> str:
-        if self.config.get("EMBEDDING_ENGINE_API_KEY") and self.config[
-            "EMBEDDING_ENGINE_MODEL"
-        ].startswith("text-embedding"):
+        # Check stored params (from constructor or config)
+        if self._api_key and self._model and self._model.startswith("text-embedding"):
             return "embedding"
         return super().id()  # default to unregistered
 
     def command(self, *args, **kwargs):
         super().command(*args, **kwargs)
         if "EMBEDDING_ENGINE_API_KEY" in kwargs:
-            openai.api_key = kwargs["EMBEDDING_ENGINE_API_KEY"]
+            self._api_key = kwargs["EMBEDDING_ENGINE_API_KEY"]
+            self._client = openai.OpenAI(api_key=self._api_key)
         if "EMBEDDING_ENGINE_MODEL" in kwargs:
             self.model = kwargs["EMBEDDING_ENGINE_MODEL"]
 
@@ -52,11 +55,11 @@ class EmbeddingEngine(Engine, OpenAIMixin):
         new_dim = kwargs.get("new_dim")
 
         try:
-            res = openai.embeddings.create(model=self.model, input=inp)
+            res = self._client.embeddings.create(model=self.model, input=inp)
         except Exception as e:
             if except_remedy is None:
                 raise e
-            callback = openai.embeddings.create
+            callback = self._client.embeddings.create
             res = except_remedy(e, inp, callback, self, *args, **kwargs)
 
         if new_dim:

{symbolicai-1.5.0 → symbolicai-1.6.0}/symai/backend/engines/neurosymbolic/__init__.py

@@ -12,6 +12,8 @@ from ...mixin import (
     OPENAI_CHAT_MODELS,
     OPENAI_REASONING_MODELS,
     OPENAI_RESPONSES_MODELS,
+    OPENROUTER_CHAT_MODELS,
+    OPENROUTER_REASONING_MODELS,
 )
 from .engine_anthropic_claudeX_chat import ClaudeXChatEngine
 from .engine_anthropic_claudeX_reasoning import ClaudeXReasoningEngine
@@ -19,6 +21,7 @@ from .engine_cerebras import CerebrasEngine
 from .engine_deepseekX_reasoning import DeepSeekXReasoningEngine
 from .engine_google_geminiX_reasoning import GeminiXReasoningEngine
 from .engine_groq import GroqEngine
+from .engine_openrouter import OpenRouterEngine
 from .engine_openai_gptX_chat import GPTXChatEngine
 from .engine_openai_gptX_reasoning import GPTXReasoningEngine
 from .engine_openai_responses import OpenAIResponsesEngine
@@ -36,6 +39,8 @@ ENGINE_MAPPING = {
     **dict.fromkeys(OPENAI_RESPONSES_MODELS, OpenAIResponsesEngine),
     **dict.fromkeys(GROQ_CHAT_MODELS, GroqEngine),
     **dict.fromkeys(GROQ_REASONING_MODELS, GroqEngine),
+    **dict.fromkeys(OPENROUTER_CHAT_MODELS, OpenRouterEngine),
+    **dict.fromkeys(OPENROUTER_REASONING_MODELS, OpenRouterEngine),
 }
 
 __all__ = [
@@ -53,6 +58,8 @@ __all__ = [
     "OPENAI_CHAT_MODELS",
     "OPENAI_REASONING_MODELS",
     "OPENAI_RESPONSES_MODELS",
+    "OPENROUTER_CHAT_MODELS",
+    "OPENROUTER_REASONING_MODELS",
     "ClaudeXChatEngine",
     "ClaudeXReasoningEngine",
     "DeepSeekXReasoningEngine",
@@ -60,5 +67,6 @@ __all__ = [
     "GPTXReasoningEngine",
     "GeminiXReasoningEngine",
     "GroqEngine",
+    "OpenRouterEngine",
     "OpenAIResponsesEngine",
 ]

{symbolicai-1.5.0 → symbolicai-1.6.0}/symai/backend/engines/neurosymbolic/engine_google_geminiX_reasoning.py

@@ -580,12 +580,25 @@ class GeminiXReasoningEngine(Engine, GoogleMixin):
     def _prepare_request_payload(self, argument):
         kwargs = argument.kwargs
 
+        stop_sequences = kwargs.get("stop")
+        if stop_sequences:
+            if isinstance(stop_sequences, str):
+                stop_sequences = [stop_sequences]
+            elif not isinstance(stop_sequences, list):
+                stop_sequences = None
+        else:
+            stop_sequences = None
+        if isinstance(stop_sequences, list):
+            stop_sequences = [seq for seq in stop_sequences if seq]
+            if not stop_sequences:
+                stop_sequences = None
+
         payload = {
             "max_output_tokens": kwargs.get("max_tokens", self.max_response_tokens),
             "temperature": kwargs.get("temperature", 1.0),
             "top_p": kwargs.get("top_p", 0.95),
             "top_k": kwargs.get("top_k", 40),
-            "stop_sequences": kwargs.get("stop", None),
+            "stop_sequences": stop_sequences,
             "stream": kwargs.get("stream", False),
         }