sqlseed-ai 0.2.5__tar.gz → 0.2.6__tar.gz

{sqlseed_ai-0.2.5 → sqlseed_ai-0.2.6}/.gitignore

@@ -59,6 +59,16 @@ dmypy.json
 *.db
 *.sqlite
 *.sqlite3
+# SQLite runtime sidecars
+*.db-wal
+*.db-shm
+*.db-journal
+*.sqlite-wal
+*.sqlite-shm
+*.sqlite-journal
+*.sqlite3-wal
+*.sqlite3-shm
+*.sqlite3-journal
 .sqlseed_cache/
 snapshots/
 .env

{sqlseed_ai-0.2.5 → sqlseed_ai-0.2.6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: sqlseed-ai
-Version: 0.2.5
+Version: 0.2.6
 Summary: Optional LLM schema analysis and contract-driven configuration repair for sqlseed
 Project-URL: Documentation, https://sunbos.github.io/sqlseed/gemma4-integration/
 Project-URL: Homepage, https://github.com/sunbos/sqlseed
@@ -46,12 +46,12 @@ backend test; installing the plugin does not perform one.
 
 ## Installation
 
-These instructions target version 0.2.5. Check [Releases](https://github.com/sunbos/sqlseed/releases)
+These instructions target version 0.2.6. Check [Releases](https://github.com/sunbos/sqlseed/releases)
 for publication status; use the source installation below to test an unpublished candidate.
 Use a Python 3.10+ virtual environment:
 
 ```bash
-python -m pip install "sqlseed-ai==0.2.5"
+python -m pip install "sqlseed-ai==0.2.6"
 ```
 
 Core 0.2.4 and older lack the shared diagnostic interfaces required by this version.
@@ -136,7 +136,7 @@ actual generated values and database constraints.
 The AI MCP entry point requires the `mcp` extra:
 
 ```bash
-python -m pip install "sqlseed-ai[mcp]==0.2.5"
+python -m pip install "sqlseed-ai[mcp]==0.2.6"
 mcp-server-sqlseed-ai
 ```
 
@@ -189,13 +189,31 @@ Strict tool calling also rejects arrays, scalars and `null` arguments as `invali
 compatibility mode may still fall back to response text. Parsed suggestions still
 require scope and rule validation.
 
-AI configuration caches include schema hashes. Schema changes invalidate cached
-suggestions; `--no-cache` bypasses them. Malformed cache metadata or configuration
-containers are treated as cache misses. Review model output before writing data.
+Single-table suggestion caches use a versioned hash of sorted column names, encoded
+as a JSON array so names containing delimiters remain distinct. The check detects
+added, removed, or renamed columns; column order, types, and constraints are not
+included. Caches using the older delimiter encoding are ignored and regenerated.
+Use `--no-cache` to analyze again without cached suggestions. This cache check is
+separate from AutoHeal's full schema fingerprint. Malformed cache metadata
+or configuration containers are treated as cache misses. Review model output before
+writing data.
+
+## Hardware estimates on macOS
+
+The AI MCP model list distinguishes Apple unified memory, Intel shared graphics
+memory, and dedicated GPU memory. Apple unified RAM is not reported as dedicated
+VRAM or added to system RAM. `unified_memory_budget_gb` is a static heuristic:
+`max(0, min(total_ram_gb * 0.75, total_ram_gb - 4))`, reserving at least 4 GiB or
+25% for the system. It is not measured free memory or a Metal allocation limit.
+
+For an identified Apple GPU on macOS, model screening compares that budget with
+both existing minimum RAM and VRAM estimates and reports at most `capable`.
+This does not verify Metal acceleration, backend/model support, or successful
+inference; loaded applications and context size can require more memory.
 
 ## Requirements
 
-These metadata requirements apply to version 0.2.5 and its source candidates.
+These metadata requirements apply to version 0.2.6 and its source candidates.
 Use local Core and plugins together when developing from source.
 
 - Python `>=3.10`

{sqlseed_ai-0.2.5 → sqlseed_ai-0.2.6}/README.md

@@ -14,12 +14,12 @@ backend test; installing the plugin does not perform one.
 
 ## Installation
 
-These instructions target version 0.2.5. Check [Releases](https://github.com/sunbos/sqlseed/releases)
+These instructions target version 0.2.6. Check [Releases](https://github.com/sunbos/sqlseed/releases)
 for publication status; use the source installation below to test an unpublished candidate.
 Use a Python 3.10+ virtual environment:
 
 ```bash
-python -m pip install "sqlseed-ai==0.2.5"
+python -m pip install "sqlseed-ai==0.2.6"
 ```
 
 Core 0.2.4 and older lack the shared diagnostic interfaces required by this version.
@@ -104,7 +104,7 @@ actual generated values and database constraints.
 The AI MCP entry point requires the `mcp` extra:
 
 ```bash
-python -m pip install "sqlseed-ai[mcp]==0.2.5"
+python -m pip install "sqlseed-ai[mcp]==0.2.6"
 mcp-server-sqlseed-ai
 ```
 
@@ -157,13 +157,31 @@ Strict tool calling also rejects arrays, scalars and `null` arguments as `invali
 compatibility mode may still fall back to response text. Parsed suggestions still
 require scope and rule validation.
 
-AI configuration caches include schema hashes. Schema changes invalidate cached
-suggestions; `--no-cache` bypasses them. Malformed cache metadata or configuration
-containers are treated as cache misses. Review model output before writing data.
+Single-table suggestion caches use a versioned hash of sorted column names, encoded
+as a JSON array so names containing delimiters remain distinct. The check detects
+added, removed, or renamed columns; column order, types, and constraints are not
+included. Caches using the older delimiter encoding are ignored and regenerated.
+Use `--no-cache` to analyze again without cached suggestions. This cache check is
+separate from AutoHeal's full schema fingerprint. Malformed cache metadata
+or configuration containers are treated as cache misses. Review model output before
+writing data.
+
+## Hardware estimates on macOS
+
+The AI MCP model list distinguishes Apple unified memory, Intel shared graphics
+memory, and dedicated GPU memory. Apple unified RAM is not reported as dedicated
+VRAM or added to system RAM. `unified_memory_budget_gb` is a static heuristic:
+`max(0, min(total_ram_gb * 0.75, total_ram_gb - 4))`, reserving at least 4 GiB or
+25% for the system. It is not measured free memory or a Metal allocation limit.
+
+For an identified Apple GPU on macOS, model screening compares that budget with
+both existing minimum RAM and VRAM estimates and reports at most `capable`.
+This does not verify Metal acceleration, backend/model support, or successful
+inference; loaded applications and context size can require more memory.
 
 ## Requirements
 
-These metadata requirements apply to version 0.2.5 and its source candidates.
+These metadata requirements apply to version 0.2.6 and its source candidates.
 Use local Core and plugins together when developing from source.
 
 - Python `>=3.10`

{sqlseed_ai-0.2.5 → sqlseed_ai-0.2.6}/README.zh-CN.md

@@ -3,7 +3,7 @@
 [English](https://github.com/sunbos/sqlseed/blob/main/plugins/sqlseed-ai/README.md) |
 **[中文](https://github.com/sunbos/sqlseed/blob/main/plugins/sqlseed-ai/README.zh-CN.md)**
 
-[sqlseed](https://sunbos.github.io/sqlseed/) 的可选 LLM Schema 分析与契约驱动配置修复插件。
+[sqlseed](https://sunbos.github.io/sqlseed/zh-CN/) 的可选 LLM Schema 分析与契约驱动配置修复插件。
 提供列规则建议、配置校验与修复，以及模板候选值生成。接受的配置可交由 Core 离线执行。
 
 支持 Google AI Studio、LM Studio、Ollama 和 OpenAI-compatible API 后端。
@@ -11,10 +11,10 @@
 
 ## 安装
 
-本文安装说明对应 0.2.5 版本；发布状态以 [Releases](https://github.com/sunbos/sqlseed/releases) 为准。测试尚未发布的候选版本时，使用下方的源码安装方式。请先创建并激活 Python 3.10+ 虚拟环境：
+本文安装说明对应 0.2.6 版本；发布状态以 [Releases](https://github.com/sunbos/sqlseed/releases) 为准。测试尚未发布的候选版本时，使用下方的源码安装方式。请先创建并激活 Python 3.10+ 虚拟环境：
 
 ```bash
-python -m pip install "sqlseed-ai==0.2.5"
+python -m pip install "sqlseed-ai==0.2.6"
 ```
 
 Core 0.2.4 及更早版本缺少本版本使用的共享诊断接口。
@@ -91,7 +91,7 @@ native/custom 方法、实际生成值及依赖数据库状态的约束需另行
 AI MCP 入口要求本包的 `mcp` extra：
 
 ```bash
-python -m pip install "sqlseed-ai[mcp]==0.2.5"
+python -m pip install "sqlseed-ai[mcp]==0.2.6"
 mcp-server-sqlseed-ai
 ```
 
@@ -194,7 +194,10 @@ Ollama 使用 JSON 对象模式。服务明确拒绝该格式时，只进行一
 
 ### 文件缓存
 
-AI 配置缓存包含 schema hash，结构变化会使旧建议失效；`--no-cache` 跳过缓存。
+单表建议缓存使用带版本标识的 schema hash，将排序后的列名编码为 JSON 数组，
+避免含分隔符的不同列名集合混淆。校验覆盖列的增删和重命名，不包含列顺序、类型或约束。
+旧分隔符编码的缓存会被忽略并重新生成；可用 `--no-cache` 跳过缓存重新分析。
+此处的缓存校验与 AutoHeal 使用的完整 schema 指纹不同。
 缓存元数据或配置容器类型无效时，按缓存未命中重新分析。
 默认路径为 macOS 的 `~/Library/Caches/sqlseed/ai_configs/`、Linux 的
 `$XDG_CACHE_HOME/sqlseed/ai_configs/`（未设置时为 `~/.cache/sqlseed/ai_configs/`），
@@ -215,9 +218,20 @@ AI 配置缓存包含 schema hash，结构变化会使旧建议失效；`--no-ca
 CLI 命令另由 `sqlseed.cli_commands` entry point 注册。本插件不实现 provider 或
 column-mapper 注册 hooks，也不要求 Core 导入 AI 实现。
 
+## macOS 硬件估算
+
+AI MCP 模型列表区分 Apple 统一内存、Intel 共享显存与独立显存。Apple 统一内存
+不会被报告为独立显存，也不会与系统 RAM 相加。`unified_memory_budget_gb` 是静态
+启发式预算：`max(0, min(total_ram_gb * 0.75, total_ram_gb - 4))`，为系统预留
+至少 4 GiB 或 25% 内存。它不是实测可用内存，也不是 Metal 分配上限。
+
+仅在 macOS 上识别出 Apple GPU 后，模型筛选才使用该预算与既有最低 RAM、VRAM
+估算同时比较，最高返回 `capable`。该结果不代表已验证 Metal 加速、后端或模型支持，
+也不保证推理成功；其他应用占用和上下文大小可能增加实际内存需求。
+
 ## 依赖
 
-以下为 0.2.5 版本及其源码候选的依赖要求。源码开发时请在同一次解析中安装本地 Core 和插件。
+以下为 0.2.6 版本及其源码候选的依赖要求。源码开发时请在同一次解析中安装本地 Core 和插件。
 
 - Python `>=3.10`
 - `sqlseed>=0.2.5.dev0,<0.3`
@@ -228,8 +242,8 @@ column-mapper 注册 hooks，也不要求 Core 导入 AI 实现。
 - 可选 `mcp` extra：`mcp>=1.0,<2`
 - 实际模型请求需要已配置且可达的后端
 
-更多信息见[AI 集成指南](https://sunbos.github.io/sqlseed/gemma4-integration.zh-CN/)、
-[升级说明](https://sunbos.github.io/sqlseed/migration.zh-CN/)和
+更多信息见[AI 集成指南](https://sunbos.github.io/sqlseed/zh-CN/gemma4-integration/)、
+[升级说明](https://sunbos.github.io/sqlseed/zh-CN/migration/)和
 [配置源码](https://github.com/sunbos/sqlseed/blob/main/plugins/sqlseed-ai/src/sqlseed_ai/config.py)。
 
 许可证：[AGPL-3.0-or-later](https://github.com/sunbos/sqlseed/blob/main/LICENSE)。

{sqlseed_ai-0.2.5 → sqlseed_ai-0.2.6}/src/sqlseed_ai/_hardware.py

@@ -11,8 +11,10 @@ from __future__ import annotations
 import ctypes
 import json
 import platform
+import re
 import subprocess
 import time
+from math import isfinite
 from typing import Any, NamedTuple
 
 from sqlseed._utils.logger import get_logger
@@ -200,8 +202,34 @@ def _detect_gpu_nvidia() -> list[dict[str, Any]]:
         return []
 
 
+def _video_memory_mb(value: object) -> int:
+    """Parse a profiler memory quantity without trusting malformed card fields."""
+    if not isinstance(value, str):
+        return 0
+    if (match := re.fullmatch(r"(\d+(?:\.\d+)?)\s*(MB|GB)", value.strip(), re.IGNORECASE)) is None:
+        return 0
+    size_mb = float(match[1]) * (1024 if match[2].upper() == "GB" else 1)
+    return int(size_mb) if isfinite(size_mb) else 0
+
+
+def _macos_gpu_memory(gpu_info: dict[str, object], vendor: str) -> tuple[str, int]:
+    """Separate shared and dedicated video memory from Apple unified RAM."""
+    if vendor == "apple":
+        return "unified", 0
+    memory_type = "shared" if "spdisplays_vram_shared" in gpu_info else "dedicated"
+    vram_mb = next(
+        (
+            size
+            for key in ("spdisplays_vram", "spdisplays_vram_shared", "_spdisplays_vram")
+            if (size := _video_memory_mb(gpu_info.get(key))) > 0
+        ),
+        0,
+    )
+    return memory_type, vram_mb
+
+
 def _detect_gpu_macos() -> list[dict[str, Any]]:
-    """Detect Apple Silicon GPU via system_profiler. macOS only."""
+    """Detect Apple, Intel and discrete GPUs via macOS system_profiler."""
     try:
         result = subprocess.run(
             ["system_profiler", "SPDisplaysDataType", "-json"],
@@ -214,26 +242,30 @@ def _detect_gpu_macos() -> list[dict[str, Any]]:
             return []
 
         data = json.loads(result.stdout)
+        if not isinstance(data, dict):
+            return []
         displays = data.get("SPDisplaysDataType", [])
+        if not isinstance(displays, list):
+            return []
         gpus: list[dict[str, Any]] = []
         for gpu_info in displays:
+            if not isinstance(gpu_info, dict):
+                continue
             name = gpu_info.get("sppci_model", "Unknown GPU")
-            vram_str = gpu_info.get("spdisplays_vram", "")
-            vram_mb = 0
-            if vram_str:
-                parts = vram_str.split()
-                if len(parts) >= 2:
-                    val = int(parts[0])
-                    unit = parts[1].upper()
-                    vram_mb = val * 1024 if "GB" in unit else val
+            if not isinstance(name, str):
+                name = "Unknown GPU"
+            vendor_label = str(gpu_info.get("spdisplays_vendor") or name).lower()
+            vendor = next((v for v in ("apple", "intel", "amd", "nvidia") if v in vendor_label), "unknown")
+            memory_type, vram_mb = _macos_gpu_memory(gpu_info, vendor)
 
             gpus.append(
                 {
                     "name": name,
                     "vram_total_mb": vram_mb,
-                    "vram_free_mb": 0,  # Apple Silicon uses unified memory; discrete VRAM is always 0
+                    "vram_free_mb": 0,  # system_profiler does not report free VRAM
                     "vram_total_gb": round(vram_mb / 1024, 1),
-                    "vendor": "apple",
+                    "vendor": vendor,
+                    "memory_type": memory_type,
                 }
             )
         return gpus
@@ -252,6 +284,13 @@ def _detect_gpus() -> list[dict[str, Any]]:
     return []
 
 
+def _has_apple_unified_memory(hw: dict[str, Any]) -> bool:
+    """Require an identified Apple GPU on macOS, including under Rosetta."""
+    return hw.get("platform", {}).get("system") == "Darwin" and any(
+        gpu.get("vendor") == "apple" and gpu.get("memory_type") == "unified" for gpu in hw.get("gpus", [])
+    )
+
+
 # ── Public API ───────────────────────────────────────────────────────
 
 
@@ -262,7 +301,8 @@ def detect_hardware() -> dict[str, Any]:
         platform: {system, release, machine}
         ram: {total_gb, available_gb}
         gpus: [{name, vram_total_mb, vram_free_mb, vram_total_gb, vendor, ...}]
-        max_vram_gb: float  (max VRAM across all GPUs, 0 if no GPU)
+        max_vram_gb: float (largest reported non-unified VRAM quantity)
+        unified_memory_budget_gb: float (heuristic budget, not measured free VRAM)
     """
     if _HardwareCache.data is not None:
         cached_time, cached_result = _HardwareCache.data
@@ -271,9 +311,9 @@ def detect_hardware() -> dict[str, Any]:
 
     ram = _detect_system_ram()
     gpus = _detect_gpus()
-    max_vram = max((g.get("vram_total_gb", 0) for g in gpus), default=0)
+    max_vram = max((g.get("vram_total_gb", 0) for g in gpus if g.get("memory_type") != "unified"), default=0)
 
-    result = {
+    result: dict[str, Any] = {
         "platform": {
             "system": platform.system(),
             "release": platform.release(),
@@ -282,7 +322,13 @@ def detect_hardware() -> dict[str, Any]:
         "ram": ram,
         "gpus": gpus,
         "max_vram_gb": max_vram,
+        "unified_memory_budget_gb": 0.0,
     }
+    if _has_apple_unified_memory(result):
+        # A static screening heuristic, not Metal's runtime allocation limit.
+        # Reserve at least 4 GiB / 25% for the OS and other applications.
+        total_ram = ram["total_gb"]
+        result["unified_memory_budget_gb"] = max(0.0, min(total_ram * 0.75, total_ram - 4.0))
 
     _HardwareCache.data = (time.monotonic(), result)
     logger.info(
@@ -340,6 +386,11 @@ def evaluate_model_status(
         return "recommended"
     if max_vram >= req.min_vram_gb:
         return "capable"
+    if _has_apple_unified_memory(hw):
+        budget = hw.get("unified_memory_budget_gb", 0)
+        # RAM and GPU allocations share this budget; never add them or grant
+        # a recommendation without verifying the backend and actual model.
+        return "capable" if budget >= max(req.min_ram_gb, req.min_vram_gb) else "insufficient"
     if total_ram >= req.min_ram_gb and max_vram == 0:
         return "cpu_only"
     if total_ram >= req.min_ram_gb:

{sqlseed_ai-0.2.5 → sqlseed_ai-0.2.6}/src/sqlseed_ai/analyzer/_caller.py

@@ -408,6 +408,8 @@ class LLMCallerMixin(_InteractionLoggingMixin):
                 error=str(e),
             )
             self._handle_llm_api_exception(e, model, streaming=False)
+        finally:
+            client.close()
 
         return self._parse_llm_completion(
             response,

{sqlseed_ai-0.2.5 → sqlseed_ai-0.2.6}/src/sqlseed_ai/analyzer/_streaming.py

@@ -8,6 +8,7 @@ request strategy (tool calling vs JSON mode vs text mode).
 from __future__ import annotations
 
 import time
+from contextlib import ExitStack
 from typing import TYPE_CHECKING, Any, NoReturn
 
 from sqlseed_ai._client import APIConnectionError, APIError, APITimeoutError, get_openai_client
@@ -217,15 +218,17 @@ class StreamingHandlerMixin(_InteractionLoggingMixin):
             raise RuntimeError("AIConfig must be initialized before calling LLM")
         client = get_openai_client(self._config)
         start_time = time.time()
-
-        if on_progress:
-            on_progress("connecting", {"model": model or self._config.model})
+        resources = ExitStack()
+        resources.callback(client.close)
 
         try:
+            if on_progress:
+                on_progress("connecting", {"model": model or self._config.model})
             kwargs = self._build_llm_kwargs(stream=True, model=model)
             kwargs["messages"] = messages
 
             stream = self._create_streaming_response(client, kwargs, strict_json=strict_json)
+            resources.callback(stream.close)
 
             content, token_count = self._collect_stream_chunks(stream, on_progress, strict_json=strict_json)
 
@@ -285,6 +288,8 @@ class StreamingHandlerMixin(_InteractionLoggingMixin):
                 error=str(e),
             )
             self._handle_llm_api_exception(e, model, streaming=True)
+        finally:
+            resources.close()
 
     def _create_streaming_response(self, client: Any, kwargs: dict[str, Any], *, strict_json: bool) -> Any:
         """Apply local JSON sampling without changing cloud streaming dispatch."""

{sqlseed_ai-0.2.5 → sqlseed_ai-0.2.6}/src/sqlseed_ai/config.py

@@ -12,6 +12,7 @@ import json
 import os
 import re
 import time
+import urllib.error
 import urllib.request
 from enum import Enum
 from typing import Any, Literal, cast
@@ -317,7 +318,9 @@ class AIConfig(BaseModel):
                 models = [str(m.get("id")) for m in data.get("data", []) if m.get("id")]
                 self._all_models_cache = (time.monotonic(), models)
                 return models
-        except (OSError, ValueError, KeyError):
+        except (OSError, ValueError, KeyError) as error:
+            if isinstance(error, urllib.error.HTTPError):
+                error.close()
             logger.debug("Could not auto-detect local models", backend=self.backend.value)
 
         # Cache negative result too, to avoid hammering a dead endpoint
@@ -626,6 +629,8 @@ class AIConfig(BaseModel):
             return info
 
         except (OSError, ValueError, RuntimeError) as e:
+            if isinstance(e, urllib.error.HTTPError):
+                e.close()
             logger.debug("Inference speed probe failed", error=str(e))
             return None
 

{sqlseed_ai-0.2.5 → sqlseed_ai-0.2.6}/src/sqlseed_ai/mcp.py

@@ -19,6 +19,7 @@ required", not "online/offline". This package requires an LLM runtime.
 from __future__ import annotations
 
 import json
+import urllib.error
 import urllib.request
 from typing import Any
 
@@ -30,7 +31,12 @@ except ImportError as _exc:  # pragma: no cover - import error path
     raise ImportError("mcp SDK not installed. Install with: pip install 'sqlseed-ai[mcp]'") from _exc
 
 from sqlseed_ai import AIBackend, AIConfig, AiConfigRefiner, AISuggestionFailedError, GemmaModel, SchemaAnalyzer
-from sqlseed_ai._hardware import MODEL_REQUIREMENTS, detect_hardware, evaluate_model_status
+from sqlseed_ai._hardware import (
+    MODEL_REQUIREMENTS,
+    _has_apple_unified_memory,
+    detect_hardware,
+    evaluate_model_status,
+)
 
 from sqlseed import ColumnConfig, DataOrchestrator
 from sqlseed._utils import paths
@@ -97,7 +103,12 @@ def sqlseed_ai_generate_yaml(
     semantics. For rule-driven offline generation, use the
     ``sqlseed_generate_yaml`` tool from ``mcp-server-sqlseed`` instead.
 
-    Supported backends: google_ai_studio (default), lm_studio, ollama, openai_compat.
+    Supported backends: google_ai_studio, lm_studio, ollama, openai_compat.
+    Backend selection uses the backend argument, then SQLSEED_AI_BACKEND,
+    then recognized URL patterns in SQLSEED_AI_BASE_URL or OPENAI_BASE_URL;
+    otherwise it defaults to openai_compat. The base_url argument only overrides
+    the endpoint and does not infer the backend. Pass backend explicitly when
+    using base_url to select a provider.
     Returns a YAML string for human review.
     """
     try:
@@ -148,7 +159,10 @@ def sqlseed_gemma4_analyze(
     table structure and recommend data generation configurations. It demonstrates
     Gemma 4's Native Function Calling feature for the AI Agent track.
 
-    Supported backends: google_ai_studio (default), lm_studio, ollama, openai_compat.
+    Supported backends: google_ai_studio, lm_studio, ollama, openai_compat.
+    Backend selection uses the backend argument, then SQLSEED_AI_BACKEND,
+    then recognized URL patterns in SQLSEED_AI_BASE_URL or OPENAI_BASE_URL;
+    otherwise it defaults to openai_compat.
     Supported models: gemma-4-26b-a4b-it (default), gemma-4-31b-it, gemma-4-12b-it,
     gemma-4-e4b-it, gemma-4-e2b-it.
     """
@@ -293,8 +307,9 @@ def _check_local_backend(backend_id: str, url: str) -> dict[str, Any]:
             data = json.loads(resp.read().decode())
             loaded = [m.get("id", "unknown") for m in data.get("data", []) if m.get("id")]
             reachable = True
-    except (OSError, ValueError):
-        pass
+    except (OSError, ValueError) as error:
+        if isinstance(error, urllib.error.HTTPError):
+            error.close()
 
     if reachable and loaded:
         reason = f"{len(loaded)} model(s) loaded"
@@ -350,12 +365,15 @@ def _build_models(hw: dict[str, Any]) -> list[dict[str, Any]]:
     for member in GemmaModel:
         status = evaluate_model_status(member.value, hw)
         req = MODEL_REQUIREMENTS.get(member.value)
+        status_description = _STATUS_ICONS.get(status, status)
+        if _has_apple_unified_memory(hw) and status != "cloud_only":
+            status_description += "; heuristic unified-memory budget, backend/model support not verified"
         models.append(
             {
                 "id": member.value,
                 "display_name": member.display_name,
                 "status": status,
-                "status_description": _STATUS_ICONS.get(status, status),
+                "status_description": status_description,
                 "local_only": member.is_local_only,
                 "requirements": {
                     "min_ram_gb": req.min_ram_gb if req else 0,
@@ -417,6 +435,7 @@ def sqlseed_list_gemma_models() -> dict[str, Any]:
             "ram": hw["ram"],
             "gpus": hw["gpus"],
             "max_vram_gb": hw["max_vram_gb"],
+            "unified_memory_budget_gb": hw.get("unified_memory_budget_gb", 0.0),
         },
     }
 

{sqlseed_ai-0.2.5 → sqlseed_ai-0.2.6}/src/sqlseed_ai/refiner.py

@@ -559,15 +559,16 @@ class AiConfigRefiner:
             table_name: Name of the table to hash.
 
         Returns:
-            Truncated SHA-256 hex digest (16 chars) of the sorted column names.
+            Versioned, truncated SHA-256 digest of unambiguously encoded sorted column names.
 
         Raises:
             ValueError: If the target has no columns to generate, including a missing table.
         """
         if not (column_names := orch.get_column_names(table_name)):
             raise ValueError(f"Table '{table_name}' does not exist or has no columns")
-        raw = "|".join(sorted(column_names))
-        return hashlib.sha256(raw.encode()).hexdigest()[:16]
+        raw = json.dumps(sorted(column_names), ensure_ascii=True, separators=(",", ":"))
+        # Keep legacy delimiter hashes separate even when a column name spells JSON.
+        return "columns-v2:" + hashlib.sha256(raw.encode("utf-8")).hexdigest()[:16]
 
     def _validate_config(
         self,