fpathlib 0.1.4.dev16__tar.gz → 0.1.6__tar.gz

{fpathlib-0.1.4.dev16/src/fpathlib.egg-info → fpathlib-0.1.6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: fpathlib
-Version: 0.1.4.dev16
+Version: 0.1.6
 Summary: A package to combine paths with metadata
 Author-email: "C. Lockhart" <clockha2@gmu.edu>
 Requires-Python: >=3.12
@@ -25,10 +25,10 @@ does this with an f-string-like path pattern -- `FPath` -- whose
 `{variable}` fields are captured out of every matching path on disk.
 
 ```python
-from fpathlib import expand_fpath
+from fpathlib import expand
 
 # Given files like data/tr1/output/0/job2.log, data/tr2/output/1/job0.log, ...
-expanded = expand_fpath("data/tr{trajectory:d}/output/{replica:d}/job{job:d}.log")
+expanded = expand("data/tr{trajectory:d}/output/{replica:d}/job{job:d}.log")
 
 for path in expanded:
     print(path, path.metadata)

{fpathlib-0.1.4.dev16 → fpathlib-0.1.6}/README.md

@@ -8,10 +8,10 @@ does this with an f-string-like path pattern -- `FPath` -- whose
 `{variable}` fields are captured out of every matching path on disk.
 
 ```python
-from fpathlib import expand_fpath
+from fpathlib import expand
 
 # Given files like data/tr1/output/0/job2.log, data/tr2/output/1/job0.log, ...
-expanded = expand_fpath("data/tr{trajectory:d}/output/{replica:d}/job{job:d}.log")
+expanded = expand("data/tr{trajectory:d}/output/{replica:d}/job{job:d}.log")
 
 for path in expanded:
     print(path, path.metadata)

{fpathlib-0.1.4.dev16 → fpathlib-0.1.6}/conda-recipe/meta.yaml

@@ -1,5 +1,5 @@
 {% set name = "fpathlib" %}
-{% set version = "0.1.3" %}
+{% set version = "0.1.5" %}
 
 package:
   name: {{ name|lower }}
@@ -7,7 +7,7 @@ package:
 
 source:
   url: https://pypi.io/packages/source/{{ name[0] }}/{{ name }}/fpathlib-{{ version }}.tar.gz
-  sha256: 059dbf937c6e965ef33f15aa214b2f354550d5dfa26b9030357954cce8c51874
+  sha256: 083299d5a41577484af5a8447d185a6ab1c190b2dfe31690375349c163ea61ab
 
 build:
   noarch: python

{fpathlib-0.1.4.dev16 → fpathlib-0.1.6}/docs/source/api.rst

@@ -23,9 +23,11 @@ FPath and ExpandedFPath
 Expanding paths
 ----------------
 
-.. autofunction:: fpathlib.expand_fpath
+.. autofunction:: fpathlib.expand
 
-.. autofunction:: fpathlib.expand_fpath_decorator
+.. autofunction:: fpathlib.iexpand
+
+.. autofunction:: fpathlib.expand_arg
 
 .. autofunction:: fpathlib.is_expandable
 

{fpathlib-0.1.4.dev16 → fpathlib-0.1.6}/docs/source/index.rst

@@ -11,10 +11,10 @@ own filename.
 
 .. code-block:: python
 
-    from fpathlib import expand_fpath
+    from fpathlib import expand
 
     # Given files like data/tr1/output/0/job2.log, data/tr2/output/1/job0.log, ...
-    expanded = expand_fpath("data/tr{trajectory:d}/output/{replica:d}/job{job:d}.log")
+    expanded = expand("data/tr{trajectory:d}/output/{replica:d}/job{job:d}.log")
 
     for path in expanded:
         print(path, path.metadata)

{fpathlib-0.1.4.dev16 → fpathlib-0.1.6}/scripts/pypi.sh

@@ -21,7 +21,14 @@ python3 -m build
 # 0.1.3/0.1.2 tag-collision incident slipped past a git-describe check even
 # though the build itself came out as a .devN), inspect what actually got
 # built. This is the real signal of whether the release is clean.
-dev_artifacts=$(ls dist/ | grep -c '\.dev[0-9]')
+#
+# `grep -c` exits 1 (not just prints "0") when it finds zero matches --
+# with `set -e` active (deploy.sh sets it, and this script is sourced into
+# that same shell), that silently killed the whole deploy right here for
+# every *clean* release build (the exact case with 0 .dev files), before
+# ever reaching twine upload. `|| true` keeps the count without letting
+# grep's "no matches" exit status abort the script.
+dev_artifacts=$(ls dist/ | grep -c '\.dev[0-9]' || true)
 if [ $allow_dev -eq 0 ] && [ "$dev_artifacts" != "0" ]
 then
   echo "built version is a dev version (tag doesn't point at a clean, distinct commit -- check 'git describe --tags --long' and 'git status'), not uploading to pypi"

{fpathlib-0.1.4.dev16 → fpathlib-0.1.6}/src/fpathlib/__init__.py

@@ -1,9 +1,9 @@
 from fpathlib.path import Path
 from fpathlib.fpath import FPath, ExpandedFPath
 from fpathlib.expand import (
-    expand_fpath,
-    iexpand_fpath,
-    expand_fpath_decorator,
+    expand,
+    iexpand,
+    expand_arg,
     is_expandable,
 )
 
@@ -11,8 +11,8 @@ __all__ = [
     "Path",
     "FPath",
     "ExpandedFPath",
-    "expand_fpath",
-    "iexpand_fpath",
-    "expand_fpath_decorator",
+    "expand",
+    "iexpand",
+    "expand_arg",
     "is_expandable",
 ]

{fpathlib-0.1.4.dev16 → fpathlib-0.1.6}/src/fpathlib/_version.py

@@ -18,7 +18,7 @@ version_tuple: tuple[int | str, ...]
 commit_id: str | None
 __commit_id__: str | None
 
-__version__ = version = '0.1.4.dev16'
-__version_tuple__ = version_tuple = (0, 1, 4, 'dev16')
+__version__ = version = '0.1.6'
+__version_tuple__ = version_tuple = (0, 1, 6)
 
-__commit_id__ = commit_id = 'g851c5dad8'
+__commit_id__ = commit_id = 'g5162f8cd0'

{fpathlib-0.1.4.dev16 → fpathlib-0.1.6}/src/fpathlib/expand.py

@@ -4,7 +4,7 @@ import parse
 from .fpath import ExpandedFPath, FPath
 
 
-def expand_fpath(fpath, *, exclude_path_patterns=None, require_metadata=True):
+def expand(fpath, *, exclude_path_patterns=None, require_metadata=True):
     """
     Use an f-string to extract out a collection of paths, where the f-string variables
     are captured and stored along the path name. This is a convenience function that
@@ -31,9 +31,9 @@ def expand_fpath(fpath, *, exclude_path_patterns=None, require_metadata=True):
     )
 
 
-def iexpand_fpath(fpath, *, exclude_path_patterns=None, require_metadata=True, errors="raise"):
+def iexpand(fpath, *, exclude_path_patterns=None, require_metadata=True, errors="raise"):
     """
-    Generator equivalent of :func:`.expand_fpath` -- lazily yields each
+    Generator equivalent of :func:`.expand` -- lazily yields each
     matching :obj:`.Path` instead of building the whole
     :obj:`.ExpandedFPath` up front. This is a convenience function that
     simply creates an :obj:`FPath` and calls its :meth:`.FPath.iexpand`
@@ -63,53 +63,51 @@ def iexpand_fpath(fpath, *, exclude_path_patterns=None, require_metadata=True, e
     )
 
 
-def expand_fpath_decorator(f=None, require_expandable=True, post_process=None):
+def expand_arg(f=None, require_expandable=False):
     """
-    Decorator for :func:`.expand_fpath`.
+    Decorator for :func:`.expand`. Expands `fpath` (if it has {}
+    named captures) before calling the wrapped function with the result;
+    otherwise calls it with `fpath` unchanged. Callers that need to react
+    differently depending on whether expansion actually happened (e.g. to
+    join in captured metadata) should check `isinstance(expanded_fpath,
+    ExpandedFPath)` themselves, inside the wrapped function -- this
+    decorator doesn't hook into that, it only handles expansion.
 
     Parameters
     ----------
     f : :obj:`callable`
-        A function that takes an :obj:`.ExpandedFPath` as its first argument.
-    post_process : :obj:`callable`
-        A function that takes the output of `f` and the :obj:`.ExpandedFPath`.
-        (Default: None).
+        A function that takes an :obj:`.ExpandedFPath` (or, if not
+        expandable and `require_expandable` is False, the original `fpath`)
+        as its first argument.
+    require_expandable : :obj:`bool`
+        Whether to require `fpath` to have {} named captures, raising
+        ValueError otherwise. If False (the default), a plain literal path
+        or glob is passed straight through to `f` unexpanded. (Default: False)
     """
 
     def decorator(f):
         @wraps(f)
         def wrapper(fpath, *args, **kwargs):
-            # If fpath is an :obj:`ExpandedFPath`, just call f with it.
+            # If fpath is already an :obj:`ExpandedFPath`, just call f with it.
             if isinstance(fpath, ExpandedFPath):
-                expanded_fpath = fpath
-                result = f(fpath, *args, **kwargs)
-
-            # Otherwise, expand fpath and call f with the result.
-            else:
-                # What happens if fpath is not expandable?
-                # If require_expandable is True, raise an error.
-                # Otherwise, just call f with the original fpath.
-                if not is_expandable(fpath):
-                    if require_expandable:
-                        msg = f"fpath not expandable: '{fpath}'"
-                        raise ValueError(msg)
-                    return f(fpath, *args, **kwargs)
-
-                # We know fpath is expandable, so we can expand it and call f
+                return f(fpath, *args, **kwargs)
+
+            # Expand fpath if it is expandable.
+            if is_expandable(fpath):
                 exclude_path_patterns = kwargs.pop("exclude_path_patterns", None)
                 require_metadata = kwargs.pop("require_metadata", True)
-                expanded_fpath = expand_fpath(
+                expanded_fpath = expand(
                     fpath,
                     exclude_path_patterns=exclude_path_patterns,
                     require_metadata=require_metadata,
                 )
-                result = f(expanded_fpath, *args, **kwargs)
-
-            # If post_process is provided, call it with the result and the expanded_fpath.
-            if post_process is not None:
-                result = post_process(result, expanded_fpath)
+                return f(expanded_fpath, *args, **kwargs)
 
-            return result
+            # fpath has no {} captures.
+            if require_expandable:
+                msg = f"fpath not expandable: '{fpath}'"
+                raise ValueError(msg)
+            return f(fpath, *args, **kwargs)
 
         return wrapper
 

{fpathlib-0.1.4.dev16 → fpathlib-0.1.6}/src/fpathlib/ext/polars.py

@@ -1,6 +1,6 @@
 from functools import wraps
 import polars as _polars
-from fpathlib import expand_fpath_decorator, ExpandedFPath
+from fpathlib import expand_arg, is_expandable, ExpandedFPath
 
 
 def __getattr__(name):
@@ -13,15 +13,59 @@ def __getattr__(name):
     return getattr(_polars, name)
 
 
-def join_metadata(df, expanded_fpath):
-    return df.join(
-        expanded_fpath.to_polars(lazy=isinstance(df, _polars.LazyFrame)),
-        on="fname",
-    )
-
+def join_metadata(f):
+    """
+    Wraps a scan_csv-shaped function `f(source, *args, **kwargs) ->
+    LazyFrame` with metadata-joining and "_fname" cleanup, applied
+    automatically to whatever `f` returns. Must be applied *inside*
+    @expand_arg (i.e. listed closer to `def`), so `source`
+    here is always the already-expanded value, not the raw caller-supplied
+    pattern:
+
+        @expand_arg
+        @join_metadata
+        def scan_csv(source, ...): ...
+
+    Guards against getting that order backwards: if `source` still looks
+    like an unexpanded {} pattern at this point, expansion can only have
+    failed to run (wrong decorator order, or this decorator used without
+    @expand_arg at all) -- raise immediately rather than
+    silently joining no metadata.
+    """
 
-@expand_fpath_decorator(post_process=join_metadata)
-def read_csv(expanded_fpath, *args, **kwargs):
+    @wraps(f)
+    def wrapper(source, *args, **kwargs):
+        lf = f(source, *args, **kwargs)
+        if not isinstance(source, ExpandedFPath) and is_expandable(source):
+            msg = (
+                f"join_metadata received an unexpanded pattern "
+                f"{source!r} -- @expand_arg must be the outer "
+                "decorator, applied above (not below) @join_metadata"
+            )
+            raise RuntimeError(msg)
+
+        # Join captured {} metadata into `lf`, keyed on the reserved
+        # "_fname" column scan_csv/scan_parquet always create, then always
+        # drop "_fname" itself. Callers that want to keep the file path
+        # under a different name must alias it to that name *before* this
+        # runs (inside `f`'s own body) -- by the time this returns,
+        # "_fname" is gone either way.
+        if isinstance(source, ExpandedFPath):
+            # "fname" is ExpandedFPath.to_polars()'s normal, public column
+            # name; rename it to the same reserved "_fname" scan_csv/
+            # scan_parquet use, so the join key can never collide with an
+            # `include_file_paths` name a caller chose (including "fname"
+            # itself).
+            metadata = source.to_polars(lazy=isinstance(lf, _polars.LazyFrame))
+            metadata = metadata.rename({"fname": "_fname"})
+            lf = lf.join(metadata, on="_fname")
+
+        return lf.drop("_fname")
+
+    return wrapper
+
+
+def read_csv(source, *args, **kwargs):
     """
     Read the paths in the collection as CSV files, and return a
     :obj:`polars.DataFrame` along with the metadata captured from the path
@@ -29,8 +73,8 @@ def read_csv(expanded_fpath, *args, **kwargs):
 
     Parameters
     ----------
-    expanded_fpath : :obj:`fpathlib.ExpandedFPath`
-        An expanded f-string path.
+    source : :obj:`fpathlib.ExpandedFPath`, or any type accepted by :obj:`polars.read_csv`
+        An expanded f-string path, or a plain path/glob.
     *args
         Positional arguments to pass to :meth:`polars.read_csv`.
     **kwargs
@@ -41,13 +85,12 @@ def read_csv(expanded_fpath, *args, **kwargs):
     :obj:`polars.DataFrame`
     """
 
-    return scan_csv.__wrapped__(expanded_fpath, *args, **kwargs).collect()
+    return scan_csv(source, *args, **kwargs).collect()
 
 
-@expand_fpath_decorator(post_process=join_metadata)
 def read_txt(
-    expanded_fpath,
-    filter_expr=None,
+    source,
+    line_filter=None,
     separator=None,
     new_columns=None,
     has_header=False,
@@ -63,10 +106,13 @@ def read_txt(
 
     Parameters
     ----------
-    expanded_fpath : :obj:`fpathlib.ExpandedFPath`
-        An expanded f-string path.
-    filter_expr : :obj:`polars.Expr`, optional
-        Filter the lines before splitting by the separator (if provided).
+    source : :obj:`fpathlib.ExpandedFPath`, or any type accepted by :obj:`polars.scan_csv`
+        An expanded f-string path, or a plain path/glob.
+    line_filter : :obj:`callable`, optional
+        A function that takes a :obj:`polars.Expr` for the line's text and
+        returns a boolean :obj:`polars.Expr`, used to filter lines before
+        splitting by the separator. E.g.
+        `lambda line: line.str.starts_with("#").not_()`.
     separator : :obj:`str`, optional
         Deliminatorg to split each line into fields.
     new_columns : :obj:`list`[:obj:`str`], optional
@@ -86,9 +132,9 @@ def read_txt(
     :obj:`polars.DataFrame`
     """
 
-    return scan_txt.__wrapped__(
-        expanded_fpath,
-        filter_expr=filter_expr,
+    return scan_txt(
+        source,
+        line_filter=line_filter,
         separator=separator,
         new_columns=new_columns,
         has_header=has_header,
@@ -97,8 +143,9 @@ def read_txt(
     ).collect()
 
 
-@expand_fpath_decorator(post_process=join_metadata)
-def scan_csv(expanded_fpath, *args, **kwargs):
+@expand_arg
+@join_metadata
+def scan_csv(source, include_file_paths=None, *args, **kwargs):
     """
     Scan the paths in the collection as CSV files, and return a
     :obj:`polars.LazyFrame` along with the metadata captured from the path
@@ -106,8 +153,12 @@ def scan_csv(expanded_fpath, *args, **kwargs):
 
     Parameters
     ----------
-    expanded_fpath : :obj:`fpathlib.ExpandedFPath`
-        An expanded f-string path.
+    source : :obj:`fpathlib.ExpandedFPath`, or any type accepted by :obj:`polars.scan_csv`
+        An expanded f-string path, or a plain path/glob.
+    include_file_paths : :obj:`str`, optional
+        Name to give a column of each row's source file path in the
+        output. The file path is always used internally to join captured
+        metadata; without this, it isn't kept in the result. (Default: None)
     *args
         Positional arguments to pass to :meth:`polars.scan_csv`.
     **kwargs
@@ -119,17 +170,19 @@ def scan_csv(expanded_fpath, *args, **kwargs):
     """
 
     lf = _polars.scan_csv(
-        expanded_fpath,
-        include_file_paths="fname",
+        source,
+        include_file_paths="_fname",
         *args,
         **kwargs,
     )
-
+    if include_file_paths is not None:
+        lf = lf.with_columns(_polars.col("_fname").alias(include_file_paths))
     return lf
 
 
-@expand_fpath_decorator(post_process=join_metadata)
-def scan_parquet(expanded_fpath, *args, **kwargs):
+@expand_arg
+@join_metadata
+def scan_parquet(source, include_file_paths=None, *args, **kwargs):
     """
     Scan the paths in the collection as a parquet file, and return a
     :obj:`polars.LazyFrame` along with the metadata captured from the path
@@ -137,8 +190,12 @@ def scan_parquet(expanded_fpath, *args, **kwargs):
 
     Parameters
     ----------
-    expanded_fpath : :obj:`fpathlib.ExpandedFPath`
-        An expanded f-string path.
+    source : :obj:`fpathlib.ExpandedFPath`, or any type accepted by :obj:`polars.scan_parquet`
+        An expanded f-string path, or a plain path/glob.
+    include_file_paths : :obj:`str`, optional
+        Name to give a column of each row's source file path in the
+        output. The file path is always used internally to join captured
+        metadata; without this, it isn't kept in the result. (Default: None)
     *args
         Positional arguments to pass to :meth:`polars.scan_parquet`.
     **kwargs
@@ -150,24 +207,25 @@ def scan_parquet(expanded_fpath, *args, **kwargs):
     """
 
     lf = _polars.scan_parquet(
-        expanded_fpath,
-        include_file_paths="fname",
+        source,
+        include_file_paths="_fname",
         *args,
         **kwargs,
     )
-
+    if include_file_paths is not None:
+        lf = lf.with_columns(_polars.col("_fname").alias(include_file_paths))
     return lf
 
 
-# TODO rename expanded_fpath as source
-@expand_fpath_decorator(require_expandable=False, post_process=join_metadata)
+@expand_arg
 def scan_txt(
-    expanded_fpath,
-    filter_expr=None,
+    source,
+    line_filter=None,
     separator=None,
     new_columns=None,
     has_header=False,
-    keep_line=False,
+    include_line=None,
+    include_file_paths=None,
     usecols=None,
     validate_schema=True,
     *args,
@@ -182,10 +240,13 @@ def scan_txt(
 
     Parameters
     ----------
-    expanded_fpath : :obj:`fpathlib.ExpandedFPath`
-        An expanded f-string path.
-    filter_expr : :obj:`polars.Expr`, optional
-        Filter the lines before splitting by the separator (if provided).
+    source : :obj:`fpathlib.ExpandedFPath`, or any type accepted by :obj:`polars.scan_csv`
+        An expanded f-string path, or a plain path/glob.
+    line_filter : :obj:`callable`, optional
+        A function that takes a :obj:`polars.Expr` for the line's text and
+        returns a boolean :obj:`polars.Expr`, used to filter lines before
+        splitting by the separator. E.g.
+        `lambda line: line.str.starts_with("#").not_()`.
     separator : :obj:`str`, optional
         Deliminatorg to split each line into fields.
     new_columns : :obj:`list`[:obj:`str`], optional
@@ -195,8 +256,13 @@ def scan_txt(
         Whether the text files have a header line that should be skipped. The header
         must have the same delimiter as the separator provided in `separator`.
         (Default: False)
-    keep_line : :obj:`bool`
-        Whether to keep the original line as a column in the output.
+    include_line : :obj:`str`, optional
+        Name to give a column of each row's original, unsplit line text in
+        the output. Without this, it isn't kept in the result. (Default: None)
+    include_file_paths : :obj:`str`, optional
+        Name to give a column of each row's source file path in the
+        output. The file path is always used internally to join captured
+        metadata; without this, it isn't kept in the result. (Default: None)
     usecols : :obj:`list`[:obj:`int`], optional
         Indexes of columns to keep in the output. If not provided, all columns are kept. Only applicable if `separator` is provided.
     validate_schema : :obj:`bool`
@@ -220,52 +286,62 @@ def scan_txt(
     :obj:`polars.LazyFrame`
     """
 
-    # TODO there are forbidden variables that should not be in expanded_fpath
-    # such as 'line' and 'fields' and 'fname'
+    # TODO there are forbidden variables that should not be in source
+    # such as '_line' and 'fields' and '_fname'
 
     # TODO schema and schema_overrides is probably broken
 
-    lf = _polars.scan_csv(
-        expanded_fpath,
-        include_file_paths="fname",
+    # Delegates to scan_csv (rather than calling _polars.scan_csv and
+    # doing the metadata-join/include_file_paths handling directly) so
+    # that handling -- including the recursive single-file sample call
+    # below, since source[0] is a plain Path, not an ExpandedFPath --
+    # isn't duplicated here. scan_csv's own @expand_arg is a
+    # no-op on `source` at this point: it's already been resolved by
+    # scan_txt's decorator, so scan_csv just sees an ExpandedFPath or an
+    # already-unexpandable literal/glob, either way with nothing left to
+    # expand.
+    lf = scan_csv(
+        source,
+        include_file_paths=include_file_paths,
         separator="\n",
-        new_columns=["line"],
+        new_columns=["_line"],
         has_header=False,
         **kwargs,
     )
 
     # Can filter lines before doing any further processing
     # This could be to remove lines with comments, etc.
-    if filter_expr is not None:
-        lf = lf.filter(filter_expr)
+    if line_filter is not None:
+        lf = lf.filter(line_filter(_polars.col("_line")))
+
+    if include_line is not None:
+        lf = lf.with_columns(_polars.col("_line").alias(include_line))
 
     # Separate lines into fields using `separator`
     if separator is not None:
         # Separate line into fields by separator
         lf = lf.with_columns(
-            _polars.col("line").str.split(separator, literal=False).alias("fields")
+            _polars.col("_line").str.split(separator, literal=False).alias("fields")
         )
-
-        if not keep_line:
-            lf = lf.drop("line")
+        lf = lf.drop("_line")
 
         # With many matched files, inferring the field count/dtypes directly
         # against the full glob is extremely slow (every file has to be opened
         # before a `.head()` takes effect). Instead, recurse on a single
-        # representative file -- expanded_fpath[0] -- and reuse its already-cheap
+        # representative file -- source[0] -- and reuse its already-cheap
         # (single-file) inference below instead of duplicating it here. Computed
         # once here since it's needed by both the field-count branch below and
         # the dtype-inference branch further down.
         infer_schema = kwargs.get("infer_schema", True)
         sample_schema = None
         if (
-            isinstance(expanded_fpath, ExpandedFPath)
-            and len(expanded_fpath) > 1
+            isinstance(source, ExpandedFPath)
+            and len(source) > 1
             and (usecols is None or infer_schema)
         ):
             sample_schema = scan_txt(
-                expanded_fpath[0],
-                filter_expr=filter_expr,
+                source[0],
+                line_filter=line_filter,
                 separator=separator,
                 new_columns=new_columns,
                 has_header=has_header,
@@ -282,7 +358,10 @@ def scan_txt(
 
         else:
             if sample_schema is not None:
-                n_fields = len(sample_schema) - 1  # minus 'fname'
+                # The recursive sample call above is always non-expandable
+                # (source[0] is a plain Path), so it already drops "_fname"
+                # itself -- no adjustment needed here.
+                n_fields = len(sample_schema)
             else:
                 n_fields = (
                     lf.head(1)
@@ -339,11 +418,9 @@ def scan_txt(
         # Infer dtypes?
         if infer_schema:
             if sample_schema is not None:
-                inferred_schema = {
-                    name: dtype
-                    for name, dtype in sample_schema.items()
-                    if name != "fname"
-                }
+                # The recursive sample call already drops "_fname" (see
+                # above), so no need to filter it back out here.
+                inferred_schema = dict(sample_schema.items())
             else:
                 sample = (
                     lf.head(kwargs.get("infer_schema_length", 100))
@@ -354,4 +431,17 @@ def scan_txt(
                 inferred_schema = _polars.read_csv(sample).schema
             lf = lf.cast(inferred_schema)
 
+    elif include_line is None:
+        # No separator means each row is just the raw line -- that's the
+        # whole point of this mode, so it has to stay visible somehow.
+        # Fall back to the traditional "line" name rather than leaking the
+        # internal "_line" name, since the caller didn't ask for a specific
+        # one via include_line.
+        lf = lf.rename({"_line": "line"})
+    else:
+        # include_line was given, so the alias was already created above
+        # (before this if/elif/else) -- drop the internal "_line" itself so
+        # it doesn't also leak into the output alongside it.
+        lf = lf.drop("_line")
+
     return lf

{fpathlib-0.1.4.dev16 → fpathlib-0.1.6/src/fpathlib.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: fpathlib
-Version: 0.1.4.dev16
+Version: 0.1.6
 Summary: A package to combine paths with metadata
 Author-email: "C. Lockhart" <clockha2@gmu.edu>
 Requires-Python: >=3.12
@@ -25,10 +25,10 @@ does this with an f-string-like path pattern -- `FPath` -- whose
 `{variable}` fields are captured out of every matching path on disk.
 
 ```python
-from fpathlib import expand_fpath
+from fpathlib import expand
 
 # Given files like data/tr1/output/0/job2.log, data/tr2/output/1/job0.log, ...
-expanded = expand_fpath("data/tr{trajectory:d}/output/{replica:d}/job{job:d}.log")
+expanded = expand("data/tr{trajectory:d}/output/{replica:d}/job{job:d}.log")
 
 for path in expanded:
     print(path, path.metadata)

fpathlib-0.1.6/src/fpathlib.egg-info/scm_version.json

@@ -0,0 +1,8 @@
+{
+  "tag": "0.1.6",
+  "distance": 0,
+  "node": "g5162f8cd0811db2b8291cbf05ff5ad278451dc81",
+  "dirty": false,
+  "branch": "main",
+  "node_date": "2026-09-21"
+}

fpathlib-0.1.4.dev16/src/fpathlib.egg-info/scm_version.json

@@ -1,8 +0,0 @@
-{
-  "tag": "0.1.3",
-  "distance": 16,
-  "node": "g851c5dad81b573754c9c1d7508faca0f7b1ab986",
-  "dirty": false,
-  "branch": "main",
-  "node_date": "2026-09-17"
-}