scihist 0.1.0__tar.gz → 0.1.9.dev0__tar.gz

{scihist-0.1.0 → scihist-0.1.9.dev0}/.gitignore

@@ -6,5 +6,11 @@ scistack-gui/extension/node_modules/
 scistack-gui/frontend/node_modules/__pycache__/ 
 *.pyc 
 __pycache__/ 
-*.pyc 
-*.pyo 
+*.pyc
+*.pyo
+
+# Generated database artifacts (DuckDB data/lineage + write-ahead logs)
+*.duckdb
+*.duckdb.wal
+*.wal
+scistack-gui/frontend/node_modules/

{scihist-0.1.0 → scihist-0.1.9.dev0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: scihist
-Version: 0.1.0
+Version: 0.1.9.dev0
 Summary: Lineage-tracked batch execution for scientific data pipelines
 Author: SciStack Contributors
 License-Expression: MIT
@@ -17,8 +17,8 @@ Classifier: Programming Language :: Python :: 3.12
 Classifier: Topic :: Scientific/Engineering
 Classifier: Typing :: Typed
 Requires-Python: >=3.9
-Requires-Dist: scidb>=0.1.0
 Requires-Dist: scilineage>=0.1.0
+Requires-Dist: scistack-db>=0.1.0
 Provides-Extra: dev
 Requires-Dist: pytest-cov>=4.0; extra == 'dev'
 Requires-Dist: pytest>=7.0; extra == 'dev'

{scihist-0.1.0 → scihist-0.1.9.dev0}/pyproject.toml

@@ -1,10 +1,14 @@
 [build-system]
-requires = ["hatchling"]
+requires = ["hatchling", "hatch-vcs"]
 build-backend = "hatchling.build"
 
+[tool.hatch.version]
+source = "vcs"
+raw-options = { search_parent_directories = true, local_scheme = "no-local-version" }
+
 [project]
 name = "scihist"
-version = "0.1.0"
+dynamic = ["version"]
 description = "Lineage-tracked batch execution for scientific data pipelines"
 readme = "README.md"
 license = "MIT"
@@ -34,7 +38,7 @@ classifiers = [
     "Typing :: Typed",
 ]
 dependencies = [
-    "scidb>=0.1.0",
+    "scistack-db>=0.1.0",
     "scilineage>=0.1.0",
 ]
 

scihist-0.1.9.dev0/src/scihist/__init__.py

@@ -0,0 +1,66 @@
+"""SciHist (deprecated) — thin shim over the consolidated scidb API.
+
+Lineage-tracked batch execution, the node-staleness API, and lineage-aware
+save now live in **scidb** (``scidb.for_each`` tracks lineage by default).
+This package remains only as a backward-compatible shim for existing imports
+(e.g. ``scistack-gui`` and the MATLAB ``+scihist`` bridge). Prefer importing
+from ``scidb`` directly; ``scihist`` will be removed in a future release.
+
+Behavioral nuances preserved by the shim:
+- ``scihist.for_each`` defaults ``skip_computed=True`` (scidb defaults False).
+- ``scihist.configure_database`` registers the DB as scilineage's cache backend.
+"""
+
+import warnings as _warnings
+
+# Core batch execution + lineage-aware save (shimmed to preserve scihist defaults)
+# Re-export DB wrappers from scidb
+# Re-export the step-function marker (replaces the removed @lineage_fcn).
+from scidb import ColumnSelection, Fixed, ForEachConfig, Merge, scistack
+
+# Re-export scifor helpers
+from scifor import Col, PathInput, get_schema, set_schema
+
+from .database import configure_database
+from .foreach import for_each, save
+from .state import (
+    check_combo_state,
+    check_multiple_nodes_state,
+    check_node_state,
+    check_pathinput_node_state,
+)
+
+_warnings.warn(
+    "scihist is deprecated; its functionality has moved to scidb. "
+    "Import from scidb instead (e.g. `from scidb import for_each, save, "
+    "configure_database`).",
+    DeprecationWarning,
+    stacklevel=2,
+)
+
+__version__ = "0.1.0"
+
+__all__ = [
+    # Core batch execution
+    "for_each",
+    "save",
+    # Configuration
+    "configure_database",
+    # Node staleness
+    "check_combo_state",
+    "check_node_state",
+    "check_pathinput_node_state",
+    "check_multiple_nodes_state",
+    # DB wrappers
+    "Fixed",
+    "Merge",
+    "ColumnSelection",
+    "ForEachConfig",
+    "PathInput",
+    # Schema helpers
+    "Col",
+    "set_schema",
+    "get_schema",
+    # Step-function marker
+    "scistack",
+]

scihist-0.1.9.dev0/src/scihist/database.py

@@ -0,0 +1,20 @@
+"""Deprecated shim — ``scihist.configure_database`` moved to scidb.
+
+Prefer importing from ``scidb`` directly. This thin wrapper remains for
+backward-compatible imports. (It previously also registered the database as
+scilineage's cache backend; the rerun cache was removed, so it now simply
+delegates to ``scidb.configure_database``.)
+"""
+
+from typing import Any
+
+
+def configure_database(
+    db_path: str,
+    schema_keys: list[str] | None = None,
+    **kwargs,
+) -> Any:
+    """Deprecated alias for ``scidb.configure_database``."""
+    from scidb import configure_database as _scidb_configure
+
+    return _scidb_configure(db_path, schema_keys)

scihist-0.1.9.dev0/src/scihist/foreach.py

@@ -0,0 +1,23 @@
+"""Deprecated shim — scihist.for_each / save moved into scidb.
+
+The lineage-tracked batch execution that lived here is now part of
+``scidb.for_each`` (which tracks lineage by default). This module remains so
+existing ``from scihist.foreach import ...`` imports keep working; prefer
+importing from ``scidb`` directly.
+
+The one behavioral nuance preserved here is ``skip_computed=True`` as the
+default (scidb.for_each defaults it to ``False``), matching the historical
+scihist.for_each contract.
+"""
+
+from scidb.foreach import for_each as _scidb_for_each
+from scidb.lineage_save import save  # noqa: F401 — re-exported for back-compat
+
+
+def for_each(*args, skip_computed: bool = True, **kwargs):
+    """Deprecated alias for ``scidb.for_each`` with ``skip_computed=True``.
+
+    scidb.for_each tracks lineage by default; this wrapper only restores the
+    historical scihist default of skipping already-computed combos.
+    """
+    return _scidb_for_each(*args, skip_computed=skip_computed, **kwargs)

scihist-0.1.9.dev0/src/scihist/state.py

@@ -0,0 +1,14 @@
+"""Deprecated shim — scihist.state moved into scidb.state.
+
+Re-exports the pipeline node staleness API from its new home in ``scidb.state``.
+Prefer importing from ``scidb`` directly.
+"""
+
+from scidb.state import (  # noqa: F401 — re-exported for back-compat
+    ComboState,
+    NodeState,
+    check_combo_state,
+    check_multiple_nodes_state,
+    check_node_state,
+    check_pathinput_node_state,
+)

scihist-0.1.0/src/scihist/__init__.py

@@ -1,68 +0,0 @@
-"""SciHist: Lineage-tracked batch execution for scientific data pipelines.
-
-This package adds lineage tracking on top of scidb. It provides the same
-for_each() interface as scidb, but automatically wraps functions in
-LineageFcn for provenance recording.
-
-Example:
-    from scihist import for_each, Fixed, configure_database
-    from scilineage import lineage_fcn
-
-    configure_database("experiment.duckdb", ["subject", "session"])
-
-    @lineage_fcn
-    def process_data(raw, calibration):
-        return raw * calibration
-
-    for_each(
-        process_data,
-        inputs={"raw": RawData, "calibration": Fixed(Calibration, session="baseline")},
-        outputs=[ProcessedData],
-        subject=[1, 2, 3],
-        session=["A", "B", "C"],
-    )
-"""
-
-from .foreach import for_each, save
-from .database import configure_database, find_by_lineage
-from .state import check_combo_state, check_node_state, check_multiple_nodes_state
-
-# Re-export DB wrappers from scidb
-from scidb import Fixed, Merge, ColumnSelection, ForEachConfig
-
-# Re-export scifor helpers
-from scifor import Col, set_schema, get_schema, PathInput
-
-# Re-export scilineage system
-from scilineage import lineage_fcn, LineageFcn, LineageFcnResult, LineageFcnInvocation
-
-__version__ = "0.1.0"
-
-__all__ = [
-    # Core batch execution
-    "for_each",
-    "save",
-    # Configuration
-    "configure_database",
-    # Lineage query
-    "find_by_lineage",
-    # Node staleness
-    "check_combo_state",
-    "check_node_state",
-    "check_multiple_nodes_state",
-    # DB wrappers
-    "Fixed",
-    "Merge",
-    "ColumnSelection",
-    "ForEachConfig",
-    "PathInput",
-    # Schema helpers
-    "Col",
-    "set_schema",
-    "get_schema",
-    # Lineage system
-    "lineage_fcn",
-    "LineageFcn",
-    "LineageFcnResult",
-    "LineageFcnInvocation",
-]

scihist-0.1.0/src/scihist/database.py

@@ -1,51 +0,0 @@
-"""SciHist database configuration — wraps scidb.configure_database with cache backend registration."""
-
-from typing import Any
-
-
-def configure_database(
-    db_path: str,
-    schema_keys: list[str] | None = None,
-    **kwargs,
-) -> Any:
-    """Configure the active database and register it as the lineage cache backend.
-
-    This is the scihist wrapper around scidb.configure_database(). In addition
-    to opening the DuckDB-backed database, it registers the database as the
-    cache backend so that lineage-tracked computations can look up previously
-    computed results.
-
-    Args:
-        db_path: Path to the DuckDB database file.
-        schema_keys: List of metadata keys that form the dataset schema.
-        **kwargs: Additional keyword arguments forwarded to scidb.configure_database().
-
-    Returns:
-        The configured DatabaseManager instance.
-    """
-    from scidb import configure_database as _scidb_configure
-    from scilineage import configure_backend
-
-    db = _scidb_configure(db_path, schema_keys, **kwargs)
-    configure_backend(db)
-    return db
-
-
-def find_by_lineage(invocation) -> list | None:
-    """Find output values by computation lineage.
-
-    Given a LineageFcnInvocation (function + inputs), finds any previously
-    computed outputs that match by querying the _lineage table via the active
-    database.
-
-    Args:
-        invocation: The LineageFcnInvocation computation to look up.
-
-    Returns:
-        List of output values if found, None otherwise.
-    """
-    from scidb.database import get_database
-
-    db = get_database()
-    lineage_hash = invocation.compute_lineage_hash()
-    return db.find_by_lineage_hash(lineage_hash)