ialdev-dataman 0.2.5__tar.gz → 0.2.7__tar.gz

{ialdev_dataman-0.2.5 → ialdev_dataman-0.2.7}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ialdev-dataman
-Version: 0.2.5
+Version: 0.2.7
 Summary: Imported as `iad.dataman` — data casting, resource models, and dataset bridge
 Author: ipcoder
 Requires-Python: >=3.10
@@ -12,10 +12,10 @@ Classifier: Programming Language :: Python :: 3
 Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
-Requires-Dist: ialdev-core>=0.2.4
-Requires-Dist: ialdev-io>=0.2.2
-Requires-Dist: ialdev-img>=0.2.2
-Requires-Dist: ialdev-maths>=0.2.3
+Requires-Dist: ialdev-core>=0.2.6
+Requires-Dist: ialdev-io>=0.2.4
+Requires-Dist: ialdev-img>=0.2.4
+Requires-Dist: ialdev-maths>=0.2.5
 Requires-Dist: pandas>=2.0.0
 Requires-Dist: numpy>=1.26.4,<2
 Requires-Dist: pydantic>=2.0

{ialdev_dataman-0.2.5 → ialdev_dataman-0.2.7}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "flit_core.buildapi"
 
 [project]
 name = "ialdev-dataman"
-version = "0.2.5"
+version = "0.2.7"
 description = "Imported as `iad.dataman` — data casting, resource models, and dataset bridge"
 readme = "README.md"
 requires-python = ">=3.10"
@@ -23,10 +23,10 @@ classifiers = [
 ]
 
 dependencies = [
-    "ialdev-core>=0.2.4",
-    "ialdev-io>=0.2.2",
-    "ialdev-img>=0.2.2",
-    "ialdev-maths>=0.2.3",
+    "ialdev-core>=0.2.6",
+    "ialdev-io>=0.2.4",
+    "ialdev-img>=0.2.4",
+    "ialdev-maths>=0.2.5",
     "pandas>=2.0.0",
     "numpy>=1.26.4,<2",
     "pydantic>=2.0",
@@ -51,6 +51,7 @@ name = "iad.dataman"
 
 [tool.pytest.ini_options]
 testpaths = ["tests"]
+addopts = "--import-mode=importlib"
 python_files = ["test_*.py"]
 python_classes = ["Test*"]
 python_functions = ["test_*"]

{ialdev_dataman-0.2.5 → ialdev_dataman-0.2.7}/src/iad/dataman/__init__.py

@@ -4,7 +4,7 @@ from .models import DataSourceRM, SchemeRM, DatasetRM, CollectionRM, DSample, di
 from .factories import create_caster, create_collection, create_sink
 from .datacast.collect import DataCollection, SinkRepo
 
-__version__ = "0.2.5"
+__version__ = "0.2.7"
 
 __all__ = [
     "__version__",

{ialdev_dataman-0.2.5 → ialdev_dataman-0.2.7}/src/iad/dataman/datacast/collect.py

@@ -741,24 +741,41 @@ class DataCollection:
         return trg
 
     def qix(self, *args, drop_level: bool = False, trans=False,
-            axis=None, key_err=True, as_dc=False, **kws) -> CollectTable | DataCollection:
-        """Fuzzy query of data collection index and return either filtered
-        version of the data collection (as_dc argument) or
-        filtered `db` table, optionally in *transformed* formed (data loaded).
-
-        Different query types are supported:
-         - by
-
-        :param args: list of values from one of the index levels
-                     - will try all until first is found or raise KeyError
+            axis=None, key_err=True, reorder=False, as_dc=False, **kws) -> CollectTable | DataCollection:
+        """Filter a data collection's index and return either a filtered
+        view of the data collection (``as_dc``) or the filtered ``db`` table,
+        optionally in *transformed* form (data loaded).
+
+        Like :meth:`DataTable.qix` this is a FILTER (values match
+        zero/one/many rows in storage order); matching is delegated to it
+        (see there for the full grammar: named levels, multi-value lists,
+        anonymous values, drop_level/keep, and opt-in reorder). For ordered
+        one-row-per-label retrieval, prefer ``select``. On top of the filter:
+
+        1. Smart query (same as db.qix):
+           ``dc.qix(view='cam1')`` / ``dc.qix(view=['cam1','cam2'], reorder=True)``
+        2. Direct index selection - pass a single list/tuple/Index to index
+           ``db`` directly (bypasses fuzzy matching; raises if any label misses):
+           ``dc.qix([('u1','cam1'), ('u2','cam2')])``
+        3. ``trans=True`` - apply column transforms so data is loaded (e.g.
+           images); returns a DataSeries of the data column.
+        4. ``as_dc=True`` (or a name str) - return a new DataCollection
+           preserving bundle/history metadata instead of a plain table.
+
+        ``trans=True`` and ``as_dc=True`` are mutually exclusive (ValueError).
+
+        :param args: anonymous values, or a single index / list-of-indices
+                     for direct selection
         :param drop_level: if True - drop found levels from the result
         :param trans: if True apply data transforms when producing results
         :param axis: if specified query only this axis
         :param key_err: if False ignore key errors
+        :param reorder: forward to DataTable.qix (order by listed values;
+                     raises on duplicates). Default False.
         :param as_dc: results are returned as a new DataCollection
         :param kws:  {level: value} - eliminates exhaustive search in all levels
 
-        :return: Transformed selection's data as DataSeries object.
+        :return: filtered db table, DataSeries (trans=True), or DataCollection (as_dc=True).
         """
         import pandas as pd
         # support of usage of direct index instead of smart queries
@@ -773,7 +790,7 @@ class DataCollection:
             if len(sub) != (len(idx) if isinstance(idx, list) else 1):
                 raise IndexError("Invalid index passed to qix")
         else:  # query based indexing
-            sub = self.db.qix(*args, drop_level=drop_level, axis=axis, key_err=key_err, **kws)
+            sub = self.db.qix(*args, drop_level=drop_level, axis=axis, key_err=key_err, reorder=reorder, **kws)
 
         if trans:
             import pandas as pd

{ialdev_dataman-0.2.5 → ialdev_dataman-0.2.7}/src/iad/dataman/datacast/transtools.py

@@ -222,16 +222,20 @@ def apply_column_transform(df: pd.Series | pd.DataFrame, *,
             except:
                 return fallback(g)
 
-    df_out[out] = df_out.apply(func, axis=1)
+    # These are deliberate writes into the working frame; suppress the
+    # false-positive SettingWithCopyWarning when the caller passed a frame
+    # that pandas flags as a copy-of-slice (e.g. a `.qix()`/`.select()` result).
+    with pd.option_context('mode.chained_assignment', None):
+        df_out.loc[:, out] = df_out.apply(func, axis=1)
 
-    if drop is True:
-        drop = set(inputs) - {out} | {trans}
-    if drop:
-        if not inplace:
-            df.drop(drop, axis=1, inplace=True)  # drop used columns in the original df, if returning new df
-            return df_out.drop(drop, axis=1)
-        df_out.drop(drop, axis=1, inplace=True)  # drop used columns in the inplace df
-    return df_out
+        if drop is True:
+            drop = set(inputs) - {out} | {trans}
+        if drop:
+            if not inplace:
+                df.drop(drop, axis=1, inplace=True)  # drop used columns in the original df, if returning new df
+                return df_out.drop(drop, axis=1)
+            df_out.drop(drop, axis=1, inplace=True)  # drop used columns in the inplace df
+        return df_out
 
 
 def row_func_prepend_transform(fnc: Callable, trans=Col.read_trans, inp=Col.path, out=Col.data):
@@ -533,7 +537,7 @@ class Fetchable:
         trans_labels = [Col.path, Col.read_trans]
 
         if self.empty:
-            self['data'] = None
+            self.loc[:, 'data'] = None
             return self
 
         if not isinstance(self, pd.DataFrame):  # Series, must be already with data
@@ -551,7 +555,10 @@ class Fetchable:
         if reuse or check_file:  # change to conditional read if required
             kws = {'check_file': check_file} | (reuse and {'check_data': reuse} or {})
             trans = '__cond_trans__'
-            self[trans] = O(cond_read_trans, **kws)
+            # Deliberate write into self (fetch documents in-place data loading);
+            # suppress the false-positive copy-of-slice warning.
+            with pd.option_context('mode.chained_assignment', None):
+                self.loc[:, trans] = O(cond_read_trans, **kws)
             trans_args = dict(drop=trans, args='series', inp=None)  # series with all columns
         else:
             trans = Col.read_trans