cfinterface 1.9.1__tar.gz → 1.10.1__tar.gz

{cfinterface-1.9.1 → cfinterface-1.10.1}/.gitignore

@@ -134,5 +134,8 @@ dmypy.json
 
 .vscode/
 
+# Implementation plans
+plans/
+
 # Build system
 uv.lock

{cfinterface-1.9.1 → cfinterface-1.10.1}/CHANGELOG.md

@@ -7,6 +7,22 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [1.10.1] - 2026-04-13
+
+### Fixed
+
+- Resolve mypy `arg-type` error in `FloatField._apply_compact` by passing the numeric value as a parameter instead of accessing `self._value` directly
+
+## [1.10.0] - 2026-04-13
+
+### Added
+
+- `compact` parameter on `FloatField` (`compact=False` by default) that omits the leading zero for values where `0 < |value| < 1`, matching Fortran fixed-width file conventions and preserving precision in narrow fields
+
+### Fixed
+
+- `RegisterReading.read()` now resets its internal data before each call, preventing register duplication when `RegisterFile.read()` retries with fallback encodings after a mid-file `UnicodeDecodeError`
+
 ## [1.9.1] - 2026-03-08
 
 ### Fixed

{cfinterface-1.9.1 → cfinterface-1.10.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: cfinterface
-Version: 1.9.1
+Version: 1.10.1
 Summary: Interface for handling custom formatted files
 Project-URL: Documentation, https://rjmalves.github.io/cfinterface/
 Project-URL: Repository, https://github.com/rjmalves/cfinterface/

{cfinterface-1.9.1 → cfinterface-1.10.1}/cfinterface/__init__.py

@@ -6,7 +6,7 @@ cfi is a Python module for handling custom formatted files
 and provide reading, storing and writing utilities.
 """
 
-__version__ = "1.9.1"
+__version__ = "1.10.1"
 
 from . import components  # noqa
 from . import data  # noqa

{cfinterface-1.9.1 → cfinterface-1.10.1}/cfinterface/components/floatfield.py

@@ -13,7 +13,7 @@ class FloatField(Field):
     by 'F' for fixed point notation and 'E' or 'D' for scientific notation.
     """
 
-    __slots__ = ["__decimal_digits", "__format", "__sep", "__type"]
+    __slots__ = ["__compact", "__decimal_digits", "__format", "__sep", "__type"]
 
     TYPES = {
         2: np.float16,
@@ -29,12 +29,14 @@ class FloatField(Field):
         format: str = "F",
         sep: str = ".",
         value: float | None = None,
+        compact: bool = False,
     ) -> None:
         super().__init__(
             size,
             starting_position,
             value,
         )
+        self.__compact = compact
         self.__decimal_digits = decimal_digits
         self.__format = format
         self.__sep = sep
@@ -63,6 +65,14 @@ class FloatField(Field):
         else:
             return np.array([self._value], dtype=self.__type).tobytes()
 
+    def _apply_compact(self, value: str, numeric: float) -> str:
+        if self.__compact and 0 < abs(numeric) < 1:
+            if value.startswith("0."):
+                return value[1:]
+            if value.startswith("-0."):
+                return "-" + value[2:]
+        return value
+
     def _textual_write(self) -> str:
         value = ""
         if self.value is not None and not _is_null(self.value):
@@ -97,6 +107,7 @@ class FloatField(Field):
                     d=self.__decimal_digits,
                     format=formatting_format,
                 ).replace("E", self.__format)
+                value = self._apply_compact(value, self.value)
                 if len(value) > self._size:
                     excess = len(value) - self._size
                     new_d = self.__decimal_digits - excess
@@ -107,6 +118,7 @@ class FloatField(Field):
                         d=new_d,
                         format=formatting_format,
                     ).replace("E", self.__format)
+                    value = self._apply_compact(value, self.value)
                     if len(value) > self._size:
                         new_d = max(0, new_d - 1)
                         value = "{:.{d}{format}}".format(
@@ -114,6 +126,7 @@ class FloatField(Field):
                             d=new_d,
                             format=formatting_format,
                         ).replace("E", self.__format)
+                        value = self._apply_compact(value, self.value)
         return value.rjust(self.size)
 
     @property

{cfinterface-1.9.1 → cfinterface-1.10.1}/cfinterface/reading/registerreading.py

@@ -82,6 +82,7 @@ class RegisterReading:
         :return: The data from the registers found in the file
         :rtype: RegisterData
         """
+        self.__data = RegisterData(DefaultRegister(data=""))
         self.__repository = factory(self.__storage)(
             content, not isfile(content), encoding
         )

{cfinterface-1.9.1 → cfinterface-1.10.1}/tests/components/test_floatfield.py

@@ -196,3 +196,108 @@ def test_floatfield_write_fuzz_equivalence():
             f"Mismatch: fmt={fmt} size={size} dec={dec} val={val} "
             f"got={result!r} expected={expected!r}"
         )
+
+
+def test_floatfield_compact_positive_small():
+    f = FloatField(5, 0, 4, format="F", value=0.0563, compact=True)
+    assert f._textual_write() == ".0563"
+
+
+def test_floatfield_compact_negative_small():
+    f = FloatField(6, 0, 4, format="F", value=-0.0563, compact=True)
+    assert f._textual_write() == "-.0563"
+
+
+def test_floatfield_compact_no_effect_above_one():
+    f = FloatField(5, 0, 2, format="F", value=1.50, compact=True)
+    assert f._textual_write() == " 1.50"
+
+
+def test_floatfield_compact_no_effect_zero():
+    f = FloatField(6, 0, 3, format="F", value=0.0, compact=True)
+    assert f._textual_write() == " 0.000"
+
+
+def test_floatfield_compact_default_false():
+    f = FloatField(5, 0, 3, format="F", value=0.0563)
+    assert f._textual_write() == "0.056"
+
+
+def test_floatfield_compact_preserves_precision():
+    """Concrete example from issue: size=5, decimal_digits=4.
+
+    Without compact: "0.0563" (6 chars) overflows → reduced to "0.056" (0.53% error).
+    With compact: ".0563" fits in 5 chars → no precision loss.
+    """
+    f_default = FloatField(5, 0, 4, format="F", value=0.0563)
+    f_compact = FloatField(5, 0, 4, format="F", value=0.0563, compact=True)
+    assert f_default._textual_write() == "0.056"
+    assert f_compact._textual_write() == ".0563"
+
+
+def test_floatfield_compact_read_roundtrip():
+    """Reading a compact-formatted value works because float('.0563') parses."""
+    f_write = FloatField(5, 0, 4, format="F", value=0.0563, compact=True)
+    written = f_write._textual_write()
+    assert written == ".0563"
+    line = written + "-rest"
+    f_read = FloatField(5, 0, 4, format="F")
+    f_read.read(line)
+    assert f_read.value == 0.0563
+
+
+def test_floatfield_compact_negative_read_roundtrip():
+    f_write = FloatField(6, 0, 4, format="F", value=-0.0563, compact=True)
+    written = f_write._textual_write()
+    assert written == "-.0563"
+    line = written + "-rest"
+    f_read = FloatField(6, 0, 4, format="F")
+    f_read.read(line)
+    assert f_read.value == -0.0563
+
+
+def test_floatfield_compact_no_effect_negative_zero():
+    f = FloatField(8, 0, 4, format="F", value=-0.0, compact=True)
+    assert f._textual_write() == " -0.0000"
+
+
+def test_floatfield_compact_overflow_still_reduces():
+    """When compact alone doesn't prevent overflow, precision is still reduced."""
+    f = FloatField(4, 0, 4, format="F", value=0.0563, compact=True)
+    result = f._textual_write()
+    assert len(result) == 4
+    assert result == ".056"
+
+
+def test_floatfield_compact_fuzz_properties():
+    """Fuzz: compact output fits field, parses correctly, is at least as precise."""
+    import random
+
+    random.seed(99)
+    for _ in range(50000):
+        size = random.randint(3, 20)
+        dec = random.randint(0, min(10, size - 1))
+        val = random.uniform(-0.999, 0.999)
+        if abs(val) < 1e-15:
+            continue
+        f_default = FloatField(size, 0, dec, format="F", value=val)
+        f_compact = FloatField(
+            size, 0, dec, format="F", value=val, compact=True
+        )
+        default_result = f_default._textual_write()
+        compact_result = f_compact._textual_write()
+        assert len(compact_result) == size, (
+            f"Wrong length: size={size} dec={dec} val={val} "
+            f"compact={compact_result!r} len={len(compact_result)}"
+        )
+        default_float = (
+            float(default_result.strip()) if default_result.strip() else 0.0
+        )
+        compact_float = (
+            float(compact_result.strip()) if compact_result.strip() else 0.0
+        )
+        # Compact must be at least as precise as default
+        assert abs(compact_float - val) <= abs(default_float - val) + 1e-15, (
+            f"Precision loss: size={size} dec={dec} val={val} "
+            f"default={default_result!r} compact={compact_result!r}"
+        )

{cfinterface-1.9.1 → cfinterface-1.10.1}/tests/files/test_registerfile.py

@@ -1,3 +1,5 @@
+import os
+import tempfile
 import warnings
 from io import StringIO
 from unittest.mock import MagicMock, patch
@@ -10,6 +12,8 @@ from cfinterface.components.literalfield import LiteralField
 from cfinterface.components.register import Register
 from cfinterface.data.registerdata import RegisterData
 from cfinterface.files.registerfile import RegisterFile
+from cfinterface.reading.registerreading import RegisterReading
+from cfinterface.storage import StorageType
 from tests.mocks.mock_open import mock_open
 
 
@@ -259,3 +263,41 @@ def test_registerfile_validate_unknown_version():
     f = VersionedRegisterFile.read(filedata)
     result = f.validate(version="v0")
     assert result.matched is False
+
+
+def test_registerreading_data_reset_between_reads():
+    """Verify that RegisterReading.__data is reset on each read() call."""
+    data = "Hello, world!"
+    filedata = DummyRegister.IDENTIFIER + " " + data + "\n"
+    reader = RegisterReading([DummyRegister], StorageType.TEXT)
+    result1 = reader.read(filedata, "utf-8")
+    assert len(result1) == 2  # root DefaultRegister + 1 DummyRegister
+    result2 = reader.read(filedata, "utf-8")
+    assert len(result2) == 2  # must be 2 again, not 4
+
+
+def test_registerfile_read_encoding_fallback_no_duplication():
+    """Verify no data duplication when first encoding fails mid-file.
+
+    A Latin-1 file with a non-UTF-8 byte causes UTF-8 decoding to fail
+    partway through. The fallback to Latin-1 must not carry over partially
+    read data from the failed UTF-8 attempt.
+    """
+    line1 = DummyRegister.IDENTIFIER + " " + "Hello, world!" + "\n"
+    line2 = DummyRegister.IDENTIFIER + " " + "with acc\u00eant! " + "\n"
+
+    class FallbackFile(RegisterFile):
+        REGISTERS = [DummyRegister]
+        ENCODING = ["utf-8", "latin-1"]
+
+    with tempfile.NamedTemporaryFile(
+        mode="wb", suffix=".txt", delete=False
+    ) as f:
+        f.write((line1 + line2).encode("latin-1"))
+        tmppath = f.name
+    try:
+        result = FallbackFile.read(tmppath)
+        # root DefaultRegister + 2 DummyRegisters = 3
+        assert len(result.data) == 3
+    finally:
+        os.unlink(tmppath)

cfinterface-1.9.1/plans/cfinterface-architecture-improvement/.implementation-state.json

@@ -1,172 +0,0 @@
-{
-  "plan": "cfinterface-architecture-improvement",
-  "last_updated": "2026-02-23T23:30:00.000000+00:00",
-  "progressive": true,
-  "tickets": {
-    "ticket-001-create-null-checking-utility.md": {
-      "status": "completed",
-      "detail_level": "detailed",
-      "agent": "python-task-automation-developer"
-    },
-    "ticket-002-replace-pd-isnull-in-fields.md": {
-      "status": "completed",
-      "detail_level": "detailed",
-      "agent": "python-task-automation-developer"
-    },
-    "ticket-003-lazy-pandas-import-in-registerfile.md": {
-      "status": "completed",
-      "detail_level": "detailed",
-      "agent": "python-task-automation-developer"
-    },
-    "ticket-004-move-pandas-to-optional-dependency.md": {
-      "status": "completed",
-      "detail_level": "detailed",
-      "agent": "python-task-automation-developer"
-    },
-    "ticket-005-compile-regex-patterns.md": {
-      "status": "completed",
-      "detail_level": "detailed",
-      "agent": "python-task-automation-developer"
-    },
-    "ticket-006-create-storagetype-enum.md": {
-      "status": "completed",
-      "detail_level": "detailed",
-      "agent": "python-task-automation-developer"
-    },
-    "ticket-007-update-factory-functions.md": {
-      "status": "completed",
-      "detail_level": "detailed",
-      "agent": "python-task-automation-developer"
-    },
-    "ticket-008-migrate-internal-storage-usage.md": {
-      "status": "completed",
-      "detail_level": "detailed",
-      "agent": "python-task-automation-developer"
-    },
-    "ticket-009-add-deprecation-warnings.md": {
-      "status": "completed",
-      "detail_level": "detailed",
-      "agent": "python-task-automation-developer"
-    },
-    "ticket-010-optimize-floatfield-write.md": {
-      "status": "completed",
-      "detail_level": "refined",
-      "agent": "python-task-automation-developer"
-    },
-    "ticket-011-add-floatfield-benchmarks.md": {
-      "status": "completed",
-      "detail_level": "refined",
-      "agent": "python-task-automation-developer"
-    },
-    "ticket-012-list-backed-registerdata.md": {
-      "status": "completed",
-      "detail_level": "refined",
-      "agent": "python-task-automation-developer"
-    },
-    "ticket-013-list-backed-blockdata.md": {
-      "status": "completed",
-      "detail_level": "refined",
-      "agent": "python-task-automation-developer"
-    },
-    "ticket-014-list-backed-sectiondata.md": {
-      "status": "completed",
-      "detail_level": "refined",
-      "agent": "python-task-automation-developer"
-    },
-    "ticket-015-computed-prev-next-properties.md": {
-      "status": "completed",
-      "detail_level": "refined",
-      "agent": "python-task-automation-developer"
-    },
-    "ticket-016-type-indexed-lookup.md": {
-      "status": "completed",
-      "detail_level": "refined",
-      "agent": "python-task-automation-developer"
-    },
-    "ticket-017-typed-field-protocol.md": {
-      "status": "completed",
-      "detail_level": "refined",
-      "agent": "python-task-automation-developer"
-    },
-    "ticket-018-typed-repository-signatures.md": {
-      "status": "completed",
-      "detail_level": "refined",
-      "agent": "python-task-automation-developer"
-    },
-    "ticket-019-typed-line-dispatch.md": {
-      "status": "completed",
-      "detail_level": "refined",
-      "agent": "python-task-automation-developer"
-    },
-    "ticket-020-typed-component-io.md": {
-      "status": "completed",
-      "detail_level": "refined",
-      "agent": "python-task-automation-developer"
-    },
-    "ticket-021-tabular-parser-engine.md": {
-      "status": "completed",
-      "detail_level": "refined",
-      "agent": "python-task-automation-developer"
-    },
-    "ticket-022-tabular-section-base.md": {
-      "status": "completed",
-      "detail_level": "refined",
-      "agent": "python-task-automation-developer"
-    },
-    "ticket-023-delimited-tabular-parsing.md": {
-      "status": "completed",
-      "detail_level": "refined",
-      "agent": "python-task-automation-developer"
-    },
-    "ticket-024-schema-version-descriptor.md": {
-      "status": "completed",
-      "detail_level": "refined",
-      "agent": "python-task-automation-developer"
-    },
-    "ticket-025-instance-level-version-binding.md": {
-      "status": "completed",
-      "detail_level": "refined",
-      "agent": "python-task-automation-developer"
-    },
-    "ticket-026-batch-read-api.md": {
-      "status": "completed",
-      "detail_level": "refined",
-      "agent": "python-task-automation-developer"
-    },
-    "ticket-027-version-validation.md": {
-      "status": "completed",
-      "detail_level": "refined",
-      "agent": "python-task-automation-developer"
-    }
-  },
-  "epics": {
-    "epic-01-remove-pandas-dependency": {
-      "phase": "completed",
-      "learning_extracted": true
-    },
-    "epic-02-compile-regex-and-storage-enum": {
-      "phase": "completed",
-      "learning_extracted": true
-    },
-    "epic-03-optimize-float-field-write": {
-      "phase": "completed",
-      "learning_extracted": true
-    },
-    "epic-04-array-backed-data-containers": {
-      "phase": "completed",
-      "learning_extracted": true
-    },
-    "epic-05-type-safe-text-binary-dispatch": {
-      "phase": "completed",
-      "learning_extracted": true
-    },
-    "epic-06-generic-tabular-parser": {
-      "phase": "completed",
-      "learning_extracted": true
-    },
-    "epic-07-schema-versioning-and-batch-ops": {
-      "phase": "completed",
-      "learning_extracted": true
-    }
-  }
-}

cfinterface-1.9.1/plans/cfinterface-architecture-improvement/00-master-plan.md

@@ -1,143 +0,0 @@
-# Master Plan: cfinterface Architecture and Performance Improvement
-
-## Executive Summary
-
-This plan addresses architectural debt and performance bottlenecks in cfinterface, a Python file schema modeling framework used to parse/write dozens of legacy HPC file formats. The improvements are ordered by impact and risk: foundational dependency removal and performance wins first (pandas removal, regex compilation, storage enum), followed by structural improvements (data containers, type-safe dispatch), and finally consumer-facing features (generic tabular parser, schema versioning). All changes maintain backward compatibility with 254+ downstream file type implementations in inewave and sintetizador-newave.
-
-## Goals & Non-Goals
-
-### Goals
-
-1. Remove the unnecessary pandas runtime dependency from cfinterface core (pandas is only used for `pd.isnull()`)
-2. Pre-compile and cache regex patterns used for identifier/block pattern matching
-3. Replace magic string `"TEXT"`/`"BINARY"` storage parameters with a type-safe enum
-4. Optimize `FloatField._textual_write()` to eliminate the O(decimal_digits) precision loop
-5. Replace linked list data containers with array-backed containers for O(1) `len()` and indexed access
-6. Introduce type-safe text/binary dispatch eliminating `Union[str, bytes]` threading
-7. Provide a generic tabular parser to reduce per-file custom DataFrame parsing in consumers
-8. Add schema versioning support and batch operations API
-
-### Non-Goals
-
-- Modifying inewave or sintetizador-newave directly (cfinterface remains independent)
-- Breaking backward compatibility without a migration path
-- Adding streaming/chunked read support (deferred to future work after array-backed containers)
-- Property accessor generation for consumer wrappers (separate future initiative)
-- Typed `Section.data` generic parameter (separate future initiative, depends on Python version constraints)
-
-## Architecture Overview
-
-### Current State
-
-```
-cfinterface v1.8.3
-  Field (base) --> FloatField, IntegerField, LiteralField, DatetimeField
-    - All import pandas for pd.isnull() null detection
-    - FloatField has O(n) textual write loop
-  Line --> composed of Fields
-    - Takes storage: str = "" parameter
-  Register/Block/Section --> contain linked-list prev/next pointers
-    - re.search() called per line without pre-compilation
-    - "TEXT"/"BINARY" magic strings threaded through all APIs
-  RegisterData/BlockData/SectionData --> linked list containers
-    - O(n) __len__(), O(n) type filtering, no indexing
-  RegisterFile/BlockFile/SectionFile --> high-level APIs
-    - RegisterFile imports pandas for _as_df() helper
-  adapters/ --> Repository pattern with factory("TEXT"/"BINARY") dispatch
-```
-
-### Target State
-
-```
-cfinterface v2.0.0
-  Field (base) --> FloatField, IntegerField, LiteralField, DatetimeField
-    - Native _is_null() helper using math.isnan/numpy, zero pandas dependency
-    - FloatField uses direct format string, no precision loop
-  StorageType enum --> TEXT, BINARY (replaces magic strings)
-  Line --> composed of Fields, uses StorageType
-  Register/Block/Section --> contain index into parent container (no linked-list pointers)
-    - Compiled regex patterns cached at class level
-  RegisterData/BlockData/SectionData --> list-backed containers
-    - O(1) len(), indexed access, type-indexed lookups
-  RegisterFile/BlockFile/SectionFile --> high-level APIs
-    - pandas import deferred to _as_df() method only (optional dependency)
-    - GenericTabularParser mixin for DataFrame construction
-    - Schema versioning with explicit contracts
-  adapters/ --> Repository pattern with StorageType enum dispatch
-```
-
-### Key Design Decisions
-
-1. **pandas becomes optional**: Move pandas to optional dependency. Use `math.isnan()` + type checking for null detection in fields. The `_as_df()` method in RegisterFile keeps pandas but imports lazily.
-2. **StorageType enum is additive**: The enum is introduced alongside the existing string API. Factory functions accept both `str` and `StorageType` during migration. The string API is deprecated but not removed.
-3. **Array-backed containers preserve API**: `RegisterData`, `BlockData`, `SectionData` switch from linked lists to `list` internally. The `previous`/`next` properties on Register/Block/Section become computed from the parent container rather than stored pointers. The public API (`append`, `preppend`, `of_type`, `get_*_of_type`) is preserved.
-4. **Regex compilation is transparent**: Patterns are compiled lazily at first use and cached as class attributes. No API change required.
-
-## Technical Approach
-
-### Tech Stack
-
-- Python >=3.10, numpy >=2.0.0
-- pandas >=2.2.3 (moved to optional dependency)
-- pytest, ruff, mypy (dev)
-
-### Component/Module Breakdown
-
-| Epic | Component                                                   | Files Modified | Risk   |
-| ---- | ----------------------------------------------------------- | -------------- | ------ |
-| 1    | Field subclasses, pyproject.toml                            | 6 files        | Low    |
-| 2    | adapters/components/repository.py, all storage string sites | 15 files       | Low    |
-| 3    | FloatField.\_textual_write()                                | 1 file         | Low    |
-| 4    | data/\*.py, Register/Block/Section prev/next                | 9 files        | Medium |
-| 5    | adapters/_, components/_, type annotations                  | 15 files       | Medium |
-| 6    | New module: cfinterface/parsers/                            | 3-5 new files  | Medium |
-| 7    | files/\*.py, new module: cfinterface/versioning/            | 5-8 files      | Medium |
-
-### Data Flow
-
-```
-File on disk
-  --> ReadingRepository (opens file, manages IO)
-    --> Reading class (dispatches lines to Register/Block/Section)
-      --> Component.read(file, storage) (reads fields from line)
-        --> Field.read(line) (extracts typed value from line slice)
-          --> Data container stores components
-            --> File class exposes data to consumer
-```
-
-### Testing Strategy
-
-- Every ticket includes unit tests for the specific change
-- Regression tests: run full existing test suite after each ticket
-- Performance benchmarks: add pytest-benchmark fixtures for regex matching, field write, data container operations
-- Integration validation: verify inewave test suite still passes (manual step at epic boundaries)
-
-## Phases & Milestones
-
-| Phase | Epic                                     | Duration  | Milestone                                                                        |
-| ----- | ---------------------------------------- | --------- | -------------------------------------------------------------------------------- |
-| 1     | Epic 1: Remove pandas dependency         | 1 week    | pandas is optional, all field tests pass without pandas installed                |
-| 1     | Epic 2: Compile regex + StorageType enum | 1-2 weeks | All re.search() calls use compiled patterns, StorageType enum in place           |
-| 2     | Epic 3: Optimize FloatField write        | 3-5 days  | FloatField.\_textual_write() uses direct formatting, benchmarks show improvement |
-| 2     | Epic 4: Array-backed data containers     | 2-3 weeks | LinkedList replaced with list, O(1) len(), existing API preserved                |
-| 3     | Epic 5: Type-safe text/binary dispatch   | 2-3 weeks | Union[str, bytes] eliminated from core APIs, adapters use StorageType            |
-| 4     | Epic 6: Generic tabular parser           | 2-3 weeks | Reusable tabular parsing for consumers, reduces boilerplate                      |
-| 4     | Epic 7: Schema versioning + batch ops    | 2-3 weeks | Version contracts, bulk read API                                                 |
-
-## Risk Analysis
-
-| Risk                                                      | Impact | Mitigation                                                                                                |
-| --------------------------------------------------------- | ------ | --------------------------------------------------------------------------------------------------------- |
-| Breaking 254+ downstream file types                       | High   | Backward-compatible changes only; deprecation before removal; string API preserved alongside enum         |
-| pandas removal breaks consumers that pass NaN values      | Medium | The \_is_null() function handles numpy NaN, float NaN, and None; comprehensive test coverage              |
-| Linked list to array migration changes iteration behavior | Medium | Preserve exact API surface; comprehensive tests for edge cases (concurrent modification during iteration) |
-| Regex compilation overhead at import time                 | Low    | Lazy compilation on first use, not at class definition time                                               |
-
-## Success Metrics
-
-1. `pytest` passes with 100% of existing tests after each epic
-2. pandas is no longer in `dependencies` (moved to `[project.optional-dependencies]`)
-3. No `re.search()` calls with raw string patterns remain
-4. `len(RegisterData)` is O(1) instead of O(n)
-5. FloatField write benchmark shows measurable improvement
-6. All `"TEXT"`/`"BINARY"` string literals replaced with `StorageType.TEXT`/`StorageType.BINARY`

cfinterface-1.9.1/plans/cfinterface-architecture-improvement/README.md

@@ -1,55 +0,0 @@
-# cfinterface Architecture and Performance Improvement
-
-## Overview
-
-Comprehensive architectural and performance improvement for the cfinterface Python package -- a file schema modeling framework used to parse/write dozens of legacy HPC file formats. Changes are ordered by impact and risk, starting with foundational improvements and progressing to consumer-facing features.
-
-## Tech Stack
-
-- Python >=3.10, numpy >=2.0.0
-- pandas >=2.2.3 (moved to optional dependency in Epic 01)
-- pytest, ruff, mypy (dev tools)
-
-## Epics
-
-| Epic | Name                             | Tickets | Detail Level | Phase     |
-| ---- | -------------------------------- | ------- | ------------ | --------- |
-| 01   | Remove pandas dependency         | 4       | Detailed     | Completed |
-| 02   | Compile regex + StorageType enum | 5       | Detailed     | Completed |
-| 03   | Optimize FloatField write        | 2       | Refined      | Completed |
-| 04   | Array-backed data containers     | 5       | Refined      | Completed |
-| 05   | Type-safe text/binary dispatch   | 4       | Refined      | Completed |
-| 06   | Generic tabular parser           | 3       | Refined      | Completed |
-| 07   | Schema versioning + batch ops    | 4       | Refined      | Completed |
-
-## Progress
-
-| Ticket     | Title                                                           | Epic    | Status    | Detail Level |
-| ---------- | --------------------------------------------------------------- | ------- | --------- | ------------ |
-| ticket-001 | Create native null-checking utility function                    | epic-01 | completed | Detailed     |
-| ticket-002 | Replace pd.isnull() in all Field subclasses                     | epic-01 | completed | Detailed     |
-| ticket-003 | Convert RegisterFile.\_as_df() to lazy pandas import            | epic-01 | completed | Detailed     |
-| ticket-004 | Move pandas to optional dependency in pyproject.toml            | epic-01 | completed | Detailed     |
-| ticket-005 | Add regex pattern compilation and caching                       | epic-02 | completed | Detailed     |
-| ticket-006 | Create StorageType enum module                                  | epic-02 | completed | Detailed     |
-| ticket-007 | Update all factory functions to accept StorageType              | epic-02 | completed | Detailed     |
-| ticket-008 | Migrate internal storage parameter usage to StorageType         | epic-02 | completed | Detailed     |
-| ticket-009 | Add deprecation warnings for string-based storage               | epic-02 | completed | Detailed     |
-| ticket-010 | Optimize FloatField fixed-point textual write                   | epic-03 | completed | Refined      |
-| ticket-011 | Add FloatField write benchmark tests                            | epic-03 | completed | Refined      |
-| ticket-012 | Implement list-backed RegisterData container                    | epic-04 | completed | Refined      |
-| ticket-013 | Implement list-backed BlockData container                       | epic-04 | completed | Refined      |
-| ticket-014 | Implement list-backed SectionData container                     | epic-04 | completed | Refined      |
-| ticket-015 | Migrate Register/Block/Section prev/next to computed properties | epic-04 | completed | Refined      |
-| ticket-016 | Add type-indexed lookup optimization                            | epic-04 | completed | Refined      |
-| ticket-017 | Design typed Field read/write protocol                          | epic-05 | completed | Refined      |
-| ticket-018 | Separate TextualRepository and BinaryRepository type signatures | epic-05 | completed | Refined      |
-| ticket-019 | Update Line class for typed dispatch                            | epic-05 | completed | Refined      |
-| ticket-020 | Update Register/Block/Section for typed IO paths                | epic-05 | completed | Refined      |
-| ticket-021 | Implement TabularParser core engine                             | epic-06 | completed | Refined      |
-| ticket-022 | Create TabularSection convenience base class                    | epic-06 | completed | Refined      |
-| ticket-023 | Add delimited tabular parsing support                           | epic-06 | completed | Refined      |
-| ticket-024 | Design SchemaVersion descriptor and version resolution utility  | epic-07 | completed | Refined      |
-| ticket-025 | Implement instance-level version binding for file classes       | epic-07 | completed | Refined      |
-| ticket-026 | Add batch read_many() API to file classes                       | epic-07 | completed | Refined      |
-| ticket-027 | Add version validation and mismatch detection                   | epic-07 | completed | Refined      |