Crate fp_index 

Row-label / index machinery for frankenpandas — every DataFrame and Series in fp-frame carries an Index that pairs labels with positional row indices, plus the alignment algebra that pandas users expect from Series + Series and DataFrame.align().

Core types

• Index: the canonical row-label container. Internally a typed Vec<IndexLabel> plus an O(1) label-lookup hashmap built lazily on first use. Pandas Index shape: monotonic probes, duplicate handling, range-style construction (Index::from_range(0..N)), name metadata.
• IndexLabel: typed label enum — Int64(i64), Float64(f64), Utf8(String), Bool(bool), Datetime64(i64), Timedelta64(i64). Lets one Index carry mixed-type labels without erasing to strings.
• MultiIndex: hierarchical multi-level index for pandas-style row MultiIndex DataFrames. Each level is itself a Vec<IndexLabel> plus an integer codes array.
• MultiIndexOrIndex: sum-type for code paths that accept either flat Index or MultiIndex.
• DuplicateKeep: enum controlling keep='first' | 'last' | False semantics in Index.duplicated / Index.drop_duplicates etc.

Alignment algebra

Binary ops between two pandas-shaped frames need to align rows by label. The aligner builds an AlignmentPlan (or MultiAlignmentPlan for N-way joins) that the caller then applies to each side’s value buffers:

• align dispatches on AlignMode (Left, Right, Inner, Outer).
• align_inner, align_left, align_union: direct single-mode entry points.
• leapfrog_union / leapfrog_intersection: N-way row alignment via a leapfrog merge over already-sorted indexes (used by multi_way_align).
• validate_alignment_plan: sanity check (lengths match, indices in bounds).

Date / time helpers

Pandas pd.date_range / pd.timedelta_range analogs:

• timedelta_range: pandas-style timedelta range builder.
• apply_date_offset / apply_date_offset_to_nanos: evaluate a DateOffset against an anchor timestamp.
• infer_freq / infer_freq_from_timestamps / infer_freq_from_nanos: pandas-style frequency inference from a sample of timestamps.
• format_datetime_ns: render a nanosecond-since-epoch i64 as the canonical YYYY-MM-DD HH:MM:SS[.f] string used in IndexLabel display and IO formatters.

Error reporting

• IndexError: structural / lookup failures (not-monotonic, not-unique, missing-label, validation-mismatch).
• TimedeltaRangeError / DateRangeError: range builder parse / step / overflow errors.

Relationship to other crates

• fp-types supplies Scalar / Timedelta / format_datetime_ns primitives.
• fp-frame stores an Index per DataFrame / Series and uses the alignment algebra here for binary ops.
• fp-join consumes alignment plans for merge-style joins.

Structs

AlignmentPlan

CategoricalIndex

Public pandas-style categorical index wrapper.

DatetimeIndex

Public pandas-style datetime index wrapper.

Index

IndexStringAccessor

MultiAlignmentPlan

Result of multi-way alignment: a union index plus per-input position vectors.

MultiIndex

A hierarchical (multi-level) index for DataFrames and Series.

PeriodFields

PeriodIndex

Public pandas-style period index wrapper.

RangeIndex

Public pandas-style range index wrapper.

TimedeltaIndex

Public pandas-style timedelta index wrapper.

Enums

AlignMode

Alignment mode for index-level join semantics.

DateOffset

A small subset of pandas pandas.tseries.offsets date offsets.

DateRangeError

Error for date_range parameter combinations.

DuplicateKeep

IndexError

IndexLabel

MultiIndexOrIndex

Result of MultiIndex::droplevel — either a MultiIndex (if 2+ levels remain) or a plain Index (if reduced to 1 level).

TimedeltaRangeError

Error for timedelta_range parameter combinations.

Functions

align

Align two indexes using the specified join mode.

align_inner

Inner alignment: only labels present in both indexes.

align_left

Left alignment: all left labels preserved, right fills with None for missing.

align_union

apply_date_offset

Apply a date offset to a timestamp string and return nanoseconds since epoch.

apply_date_offset_to_nanos

Apply a date offset to a nanosecond timestamp.

bdate_range

Create a DatetimeIndex with default weekday-only business-day values.

date_range

Create a DatetimeIndex with evenly spaced values.

format_datetime_ns

infer_freq

Infer a pandas-style frequency string from a DatetimeIndex.

infer_freq_from_nanos

Infer a pandas-style frequency string from nanosecond timestamps.

infer_freq_from_timestamps

Infer a pandas-style frequency string from timestamp strings.

leapfrog_intersection

Leapfrog intersection: labels present in ALL input indexes.

leapfrog_union

K-way merge union of multiple sorted iterators.

multi_way_align

Multi-way alignment: union all indexes, then compute position vectors.

timedelta_range

Create a TimedeltaIndex with evenly spaced values.

validate_alignment_plan