hyperforge 1.0.0.post403__tar.gz → 1.0.0.post414__tar.gz

{hyperforge-1.0.0.post403 → hyperforge-1.0.0.post414}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: hyperforge
-Version: 1.0.0.post403
+Version: 1.0.0.post414
 Summary: Agentic Framework for Orchestrated Runtime, Governance, and Execution
 Author: AI Data Team
 Author-email: AI Data Team <learning@nuclia.com>

{hyperforge-1.0.0.post403 → hyperforge-1.0.0.post414}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "uv_build"
 
 [project]
 name = "hyperforge"
-version = "1.0.0.post403"
+version = "1.0.0.post414"
 description = "Agentic Framework for Orchestrated Runtime, Governance, and Execution"
 readme = "README.md"
 requires-python = ">=3.12"

{hyperforge-1.0.0.post403 → hyperforge-1.0.0.post414}/pyproject.toml.orig

@@ -4,7 +4,7 @@ build-backend = "uv_build"
 
 [project]
 name = "hyperforge"
-version = "1.0.0.post403"
+version = "1.0.0.post414"
 authors = [{ name = "AI Data Team", email = "learning@nuclia.com" }]
 description = "Agentic Framework for Orchestrated Runtime, Governance, and Execution"
 readme = "README.md"

hyperforge-1.0.0.post414/src/hyperforge/alembic/versions/8c7d2e4f91ab_structure_legacy_llm_fields.py

@@ -0,0 +1,24 @@
+"""Structure legacy string model fields in stored agent configurations.
+
+Revision ID: 8c7d2e4f91ab
+Revises: 6a23d8ef74c1
+"""
+
+from typing import Sequence, Union
+
+from alembic import op
+
+from hyperforge.db.legacy_llm_migration import migrate_legacy_llm_fields
+
+revision: str = "8c7d2e4f91ab"
+down_revision: Union[str, None] = "6a23d8ef74c1"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    migrate_legacy_llm_fields(op.get_bind())
+
+
+def downgrade() -> None:
+    pass

hyperforge-1.0.0.post414/src/hyperforge/db/legacy_llm_migration.py

@@ -0,0 +1,132 @@
+"""Convert old model strings in stored Hyperforge agent configs to LLMConfig.
+
+Only fields already declared as LLMField in Hyperforge are included. Agent
+plugins with string-typed model fields must remain strings until their schemas
+and runtimes are migrated separately.
+"""
+
+import json
+from typing import Any
+
+from sqlalchemy import text
+from sqlalchemy.engine import Connection
+
+from hyperforge.db.llm_migration_utils import HYPERFORGE_LLM_TABLES
+from hyperforge.llm_config import LLM_CONFIG_TYPE
+
+# Keep this inventory aligned with the LLMField declarations in Hyperforge.
+# Context fields are inherited by ContextAgentConfig subclasses, including
+# the installed progress-agents plugins. Their own string-typed model fields
+# are deliberately excluded from MODEL_FIELDS.
+CONTEXT_MODULES = {
+    "a2a",
+    "advanced_ask",
+    "ask",
+    "basic_ask",
+    "brave",
+    "context_conditional",
+    "cypher",
+    "google",
+    "http",
+    "marklogic",
+    "mcp",
+    "multi_mcp",
+    "pandas",
+    "perplexity",
+    "perplexity_search",
+    "restricted",
+    "salesforce",
+    "smart",
+    "snowflake",
+    "sparql",
+    "sql",
+    "static",
+    "static_string",
+    "sitefinity",
+    "sync",
+}
+CONTEXT_FIELDS = {"context_validation_model", "rephrase_model", "summarize_model"}
+MODEL_FIELDS = {
+    "advanced_ask": {"generative_model"},
+    "ask": {"configuration_model"},
+    "basic_ask": {"generative_model"},
+    "context_conditional": {"model"},
+    "external": {"model"},
+    "generate": {"model"},
+    "generation_conditional": {"model"},
+    "mcp": {"tool_choice_model", "sampling_model"},
+    "multi_mcp": {"summarize_model", "tool_choice_model"},
+    "post_conditional": {"model"},
+    "pre_conditional": {"model"},
+    "related": {"model"},
+    "rephrase": {"model"},
+    "restart": {"model"},
+    "restricted": {"decision_model"},
+    "salesforce": {"tool_choice_model", "sampling_model"},
+    "smart": {"planner_model", "executor_model"},
+    "summarize": {"model"},
+    "sync": {"generative_model"},
+}
+NESTED_FIELDS_BY_MODULE = {
+    "advanced_generation": {"summarize_config", "data_viz_config"},
+    "context_conditional": {"then", "else_"},
+    "generation_conditional": {"then", "else_"},
+    "multi_mcp": {"configs"},
+    "post_conditional": {"then", "else_"},
+    "pre_conditional": {"then", "else_"},
+    "restricted": {"agents"},
+    "smart": {"registered_agents"},
+}
+
+
+def wrap_legacy_llm_fields(config: Any) -> bool:
+    """Wrap string-valued LLMFields, including those inside nested agents."""
+    changed = False
+    if isinstance(config, list):
+        for item in config:
+            changed = wrap_legacy_llm_fields(item) or changed
+    elif isinstance(config, dict) and isinstance(config.get("module"), str):
+        module = config["module"]
+        fields = MODEL_FIELDS.get(module, set())
+        if module in CONTEXT_MODULES:
+            fields = fields | CONTEXT_FIELDS
+        for field in fields:
+            value = config.get(field)
+            if isinstance(value, str):
+                config[field] = {"_type": LLM_CONFIG_TYPE, "model_id": value}
+                changed = True
+        nested_fields = NESTED_FIELDS_BY_MODULE.get(module, set())
+        if module in CONTEXT_MODULES:
+            nested_fields = nested_fields | {"fallback", "next_agent"}
+        for field in nested_fields:
+            if field in config:
+                changed = wrap_legacy_llm_fields(config[field]) or changed
+    return changed
+
+
+def migrate_legacy_llm_fields(conn: Connection) -> int:
+    """Convert stored agent configs, leaving unrelated and structured data alone."""
+    modified = 0
+    for table, column in HYPERFORGE_LLM_TABLES:
+        last_id = "00000000-0000-0000-0000-000000000000"
+        while rows := conn.execute(
+            text(
+                f"SELECT id, {column} FROM {table} "
+                "WHERE id > CAST(:last_id AS uuid) ORDER BY id LIMIT 500"
+            ),
+            {"last_id": last_id},
+        ).fetchall():
+            for row_id, config in rows:
+                if isinstance(config, str):
+                    config = json.loads(config)
+                if wrap_legacy_llm_fields(config):
+                    conn.execute(
+                        text(
+                            f"UPDATE {table} SET {column} = CAST(:config AS jsonb) "
+                            "WHERE id = :id"
+                        ),
+                        {"config": json.dumps(config), "id": row_id},
+                    )
+                    modified += 1
+            last_id = str(rows[-1][0])
+    return modified

{hyperforge-1.0.0.post403 → hyperforge-1.0.0.post414}/src/hyperforge/llm_config.py

@@ -165,13 +165,13 @@ class LLMDefaults(BaseSettings):
     model_config = SettingsConfigDict(env_prefix="HYPERFORGE_LLM_")
 
     # General purpose (low cost, fast)
-    default: str = "chatgpt-azure-4o-mini"
+    default: str = "chatgpt-azure-5.6-luna"
     # Smart tier (complex planning, tool use)
-    smart: str = "chatgpt-4.1"
+    smart: str = "chatgpt-azure-5.6-terra"
     # Fast tier (high throughput, configuration)
-    fast: str = "gemini-2.5-flash"
+    fast: str = "gemini-3.6-flash"
     # Reasoning tier (decision making, conditions)
-    reasoning: str = "chatgpt-o3-mini"
+    reasoning: str = "chatgpt-azure-5.6-luna"
 
 
 llm_defaults = LLMDefaults()