python-jack-knife 0.7.0__tar.gz → 0.7.1__tar.gz

{python_jack_knife-0.7.0/src/python_jack_knife.egg-info → python_jack_knife-0.7.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: python-jack-knife
-Version: 0.7.0
+Version: 0.7.1
 Summary: Python Jack Knife – a command line data processor
 Author-email: Mike Schultz <mike.schultz@gmail.com>
 License: 

{python_jack_knife-0.7.0 → python_jack_knife-0.7.1}/src/pjk/history.py

@@ -30,6 +30,9 @@ def read_history(log_path: str) -> List[int]:
                 if not line:
                     continue
 
+                if 'pjk ' in line: # legacy
+                    line = line.split('pjk ', 1)[1]
+
                 # Expected format: <command_string>
                 line = line.strip()
                         

{python_jack_knife-0.7.0 → python_jack_knife-0.7.1}/src/pjk/integrations/postgres_pipe.py

@@ -35,9 +35,12 @@ class DBClient:
         self.conn = DBClient._connection
 
     def close(self):
+        import pg8000 # lazy
         if self.conn is not None:
             try:
                 self.conn.close()
+            except pg8000.exceptions.InterfaceError:
+                pass
             finally:
                 DBClient._connection = None
 
@@ -123,10 +126,15 @@ class PostgresPipe(QueryPipe,Integration):
 
         self.params_field = "params"  # optional: list/tuple (positional) or dict (named)
 
+        self.client = None
+
     def reset(self):
         # stateless across reset
         pass
 
+    def close(self):
+        self.client.close()
+
     def _make_header(self, cur, query: str, params=None) -> Dict[str, Any]:
         """
         Inspect the cursor and build a full header record.
@@ -155,7 +163,7 @@ class PostgresPipe(QueryPipe,Integration):
         return h
 
     def execute_query_returning_S_xO_iterable(self, record):
-        client = DBClient(
+        self.client = DBClient(
             host=self.db_host,
             username=self.db_user,
             password=self.db_pass,
@@ -173,7 +181,7 @@ class PostgresPipe(QueryPipe,Integration):
             params = record.get(self.params_field)          # single-exec params
             batch  = record.get("batch_params", None)       # list[tuple|dict] for batching
 
-            cur = client.conn.cursor()
+            cur = self.client.conn.cursor()
             try:
                 did_executemany = False
 
@@ -215,4 +223,5 @@ class PostgresPipe(QueryPipe,Integration):
             finally:
                 cur.close()
         finally:
-            client.close()
+            pass
+        #    client.close()

{python_jack_knife-0.7.0 → python_jack_knife-0.7.1}/src/pjk/pipes/factory.py

@@ -38,7 +38,7 @@ COMPONENTS = {
         'reduce': ReducePipe,        
         'sort': SortPipe,
         'where': WherePipe,
-        'sel': SelectFields,
+        'select': SelectFields,
         'sample': SamplePipe,
         'explode': DenormPipe,
         'postgres': PostgresPipe,

{python_jack_knife-0.7.0 → python_jack_knife-0.7.1}/src/pjk/pipes/select.py

@@ -10,12 +10,12 @@ class SelectFields(DeepCopyPipe):
     @classmethod
     def usage(cls):
         usage = Usage(
-            name='sel',
+            name='select',
             desc='Select specific fields from each record.',
             component_class=cls
         )
         usage.def_arg(name='fields', usage='Comma-separated list of fields to retain')
-        usage.def_example(expr_tokens=["{id:1, dir:'up', color:'blue'}", 'sel:id,color'], expect="id: 1, color:'blue'")
+        usage.def_example(expr_tokens=["{id:1, dir:'up', color:'blue'}", 'select:id,color'], expect="id: 1, color:'blue'")
         return usage
 
     def __init__(self, ptok: ParsedToken, usage: Usage):

{python_jack_knife-0.7.0 → python_jack_knife-0.7.1}/src/pjk/sinks/graph_bar_line.py

@@ -20,10 +20,6 @@ from typing import Any, Dict, Iterable, List, Optional, Sequence
 from datetime import date, datetime
 from collections import defaultdict
 
-import numpy as np
-import pandas as pd
-
-
 # ----------------------------- Public Params -----------------------------
 @dataclass
 class GraphParams:
@@ -48,6 +44,8 @@ class TimeDetector:
 
     @staticmethod
     def is_time(xs: pd.Series) -> bool:
+        import numpy as np # lazy
+        import pandas as pd # lazy
         # Already datetime dtype?
         if pd.api.types.is_datetime64_any_dtype(xs):
             return True
@@ -74,6 +72,7 @@ class TimeDetector:
 
     @staticmethod
     def parse_times(series: pd.Series) -> pd.Series:
+        import pandas as pd # lazy
         numeric = pd.to_numeric(series, errors="coerce")
         parsed = None
         if numeric.notna().mean() >= 0.9:
@@ -92,6 +91,7 @@ class MultiYAdapter:
     """Builds wide dataframe: columns = ['x'] + y_fields; sums duplicates of x."""
     @staticmethod
     def to_df(records: Iterable[Dict[str, Any]], x_field: str, y_fields: Sequence[str]) -> pd.DataFrame:
+        import pandas as pd # lazy
         rows: List[Dict[str, Any]] = []
         for r in records:
             if x_field not in r:
@@ -137,12 +137,14 @@ class SingleYWithSetsAdapter:
 # ----------------------------- Plotter -----------------------------
 class GraphPlotter:
     def __init__(self, params: GraphParams):
+        import numpy as np
         self.pms = params
         self.y_fields = list(dict.fromkeys(self.pms.y_fields))  # dedupe, preserve order
 
     def plot(self, chart_type: str = "line"):
-        import matplotlib.pyplot as plt
-        import matplotlib.dates as mdates
+        import matplotlib.pyplot as plt # lazy
+        import matplotlib.dates as mdates # lazy
+        import pandas as pd # lazy
 
         fig = plt.figure()
         ax = plt.gca()
@@ -258,7 +260,7 @@ class GraphPlotter:
     # ---------- Formatting helpers ----------
     @staticmethod
     def _format_time_axis(ax, df: pd.DataFrame) -> None:
-        import matplotlib.dates as mdates
+        import matplotlib.dates as mdates # lazy
         fig = ax.get_figure()
         ts = df["ts"]
         if ts.empty:
@@ -322,7 +324,7 @@ class GraphPlotter:
 
     # ---------- Misc ----------
     def _apply_args_dict(self) -> None:
-        import matplotlib.pyplot as plt
+        import matplotlib.pyplot as plt # lazy
         for name, val in getattr(self.pms, "args_dict", {}).items():
             fn = getattr(plt, name, None)
             if callable(fn):
@@ -345,7 +347,7 @@ def graph_bar_line(obj, type):
     Returns (fig, ax) for optional downstream tweaks (safe to ignore).
     """
     # Lazy import (ensures MPL backend)
-    import matplotlib.pyplot as plt  # noqa: F401
+    import matplotlib.pyplot as plt  # noqa: F401 # lazy
 
     # Normalize y_fields from string or list
     raw_y = obj.y_field if isinstance(obj.y_field, str) else str(obj.y_field)

{python_jack_knife-0.7.0 → python_jack_knife-0.7.1}/src/pjk/sources/npy_source.py

@@ -4,18 +4,15 @@
 import json
 from typing import Iterator, Dict, Any
 
-import numpy as np
-from pjk.usage import NoBindUsage
-from pjk.components import Source
 from pjk.sources.lazy_file import LazyFile
 from pjk.sources.format_source import FormatSource
 from pjk.log import logger
 
-
 class NpySource(FormatSource):
     extension = 'npy'
 
     def __init__(self, lazy_file: LazyFile):
+        super().__init__(root=None)
         self.lazy_file = lazy_file
         self.num_vecs = 0
 
@@ -32,9 +29,11 @@ class NpySource(FormatSource):
 
         try:
             # Use mmap to avoid loading entire array in RAM at once.
+            import numpy as np #lazy import
             arr = np.load(path, mmap_mode="r", allow_pickle=False)
         except Exception as e:
             logger.error(f"Failed to load .npy file at {path}: {e}")
+            raise Exception(f"Failed to load .npy file at {path}: {e}")
             return
 
         if arr.size == 0:

{python_jack_knife-0.7.0 → python_jack_knife-0.7.1}/src/pjk/version.py

@@ -1,4 +1,4 @@
 # SPDX-License-Identifier: Apache-2.0
 # Copyright 2024 Mike Schultz
 
-__version__ = "0.7.0"
+__version__ = "0.7.1"

{python_jack_knife-0.7.0 → python_jack_knife-0.7.1/src/python_jack_knife.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: python-jack-knife
-Version: 0.7.0
+Version: 0.7.1
 Summary: Python Jack Knife – a command line data processor
 Author-email: Mike Schultz <mike.schultz@gmail.com>
 License: