sinlib 0.2.0__tar.gz → 0.2.2__tar.gz

{sinlib-0.2.0 → sinlib-0.2.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sinlib
-Version: 0.2.0
+Version: 0.2.2
 Summary: Sinhala NLP Toolkit
 Project-URL: Code, https://github.com/Ransaka/sinlib
 Project-URL: Docs, https://sinlib.readthedocs.io
@@ -35,6 +35,7 @@ Classifier: Programming Language :: Python :: 3.11
 Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
 Requires-Python: >=3.9.7
 Requires-Dist: huggingface-hub
+Requires-Dist: matplotlib>=3.9.4
 Requires-Dist: numpy>=1.24.0
 Requires-Dist: tqdm>=4.64.1
 Requires-Dist: transformers>=4.31.0

{sinlib-0.2.0 → sinlib-0.2.2}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "sinlib"
-version = "0.2.0"
+version = "0.2.2"
 description = "Sinhala NLP Toolkit"
 authors = [
     { name = "Ransaka", email = "ransaka.ravihara@gmail.com" }
@@ -9,7 +9,8 @@ dependencies = [
     "numpy >= 1.24.0",
     "transformers >= 4.31.0",
     "tqdm >= 4.64.1",
-    "huggingface_hub"
+    "huggingface_hub",
+    "matplotlib>=3.9.4",
 ]
 license = { file = "LICENSE" }
 readme = "README.md"

{sinlib-0.2.0 → sinlib-0.2.2}/src/sinlib/__init__.py

@@ -2,10 +2,10 @@
 Sinlib: A comprehensive library for Sinhala text processing.
 
 Provides tools for tokenization, spell-checking, romanization, and
-transliteration of Sinhala text, along with preprocessing utilities.
+transliteration of Sinhala text, along with preprocessing and visualization utilities.
 
-Available Classes
------------------
+Available Classes & Functions
+-----------------------------
 Tokenizer
     Character-level tokenizer for Sinhala text. Mirrors the HuggingFace
     ``PreTrainedTokenizer`` interface — use ``Tokenizer.from_pretrained()``
@@ -23,6 +23,9 @@ preprocessing
     Utility module exposing ``remove_english_characters``,
     ``remove_non_printable``, and ``get_sinhala_character_ratio``.
 
+setup_matplotlib
+    Configures and returns a FontProperties object for Matplotlib rendering.
+
 Romanizer
     Converts Sinhala text to Roman (Latin) script.
 
@@ -39,7 +42,7 @@ from sinlib.subword import SubwordTokenizer
 from sinlib.spellcheck import TypoDetector
 from sinlib.romanize import Romanizer
 from sinlib.transliterate import Transliterator  # noqa: F401
-from sinlib.utils import preprocessing
+from sinlib.utils import preprocessing, setup_matplotlib
 
 __all__: List[str] = [
     "BatchEncoding",
@@ -48,6 +51,7 @@ __all__: List[str] = [
     "TypoDetector",
     "Romanizer",
     "preprocessing",
+    "setup_matplotlib",
 ]
 
-__version__ = "0.1.13"
+__version__ = "0.2.2"

{sinlib-0.2.0 → sinlib-0.2.2}/src/sinlib/utils/__init__.py

@@ -4,10 +4,12 @@ from .preprocessing import (
     get_sinhala_character_ratio,
     normalize_sinhala,
 )
+from .visualize import setup_matplotlib
 
 __all__ = [
     "remove_english_characters",
     "remove_non_printable",
     "get_sinhala_character_ratio",
     "normalize_sinhala",
+    "setup_matplotlib",
 ]

{sinlib-0.2.0 → sinlib-0.2.2}/src/sinlib/utils/preprocessing.py

@@ -404,6 +404,7 @@ def get_sinhala_character_ratio(
             with multiprocessing.Pool() as pool:
                 results = pool.map(partial_process_text, text)
 
+        encodings = [[tok for tok in res[0] if tok != " "] for res in results]
         sinhala_lengths = [tok[1] for tok in results]
         return [
             (l / len(enc)) if enc else 0.0

sinlib-0.2.2/src/sinlib/utils/visualize.py

@@ -0,0 +1,32 @@
+import matplotlib.font_manager as fm
+from huggingface_hub import hf_hub_download
+
+
+def setup_matplotlib() -> fm.FontProperties:
+    """Download Noto Sans Sinhala font and return a Matplotlib FontProperties object.
+
+    This font supports both Latin (English) and Sinhala Unicode blocks natively,
+    allowing mixed-script strings to be rendered cleanly in a single text element.
+
+    Returns:
+        matplotlib.font_manager.FontProperties bound to the cached Noto Sans Sinhala font.
+
+    Example:
+        >>> import matplotlib.pyplot as plt
+        >>> from sinlib import setup_matplotlib
+        >>> fp = setup_matplotlib()
+        >>> plt.text(0.5, 0.5, "Target Token: 'සිංහල'", fontproperties=fp)
+    """
+    try:
+        font_path = hf_hub_download(
+            repo_id="Ransaka/sinlib",
+            filename="fonts/NotoSansSinhala-Regular.ttf",
+            repo_type="model",
+        )
+    except Exception as e:
+        raise RuntimeError(
+            "Could not download Noto Sans Sinhala font from Hugging Face Hub. "
+            f"Error: {e}"
+        )
+
+    return fm.FontProperties(fname=font_path)

{sinlib-0.2.0 → sinlib-0.2.2}/tests/test_phase3_enhancements.py

@@ -124,4 +124,22 @@ def test_normalize_sinhala():
     assert normalize_sinhala("කේ") == "කේ"
 
 
-
+def test_get_sinhala_character_ratio():
+    from sinlib.utils.preprocessing import get_sinhala_character_ratio
+    
+    # 1. Single string tests
+    assert get_sinhala_character_ratio("මම ගෙදර ගියා.") == 1.0
+    assert get_sinhala_character_ratio("Hello") == 0.0
+    
+    # 2. List < 10 tests
+    ratios_small = get_sinhala_character_ratio(["මම ගෙදර ගියා.", "Hello"])
+    assert len(ratios_small) == 2
+    assert ratios_small[0] == 1.0
+    assert ratios_small[1] == 0.0
+    
+    # 3. List >= 10 tests (triggers multiprocessing pool path)
+    large_list = ["මම ගෙදර ගියා."] * 5 + ["Hello"] * 5
+    ratios_large = get_sinhala_character_ratio(large_list)
+    assert len(ratios_large) == 10
+    assert all(r == 1.0 for r in ratios_large[:5])
+    assert all(r == 0.0 for r in ratios_large[5:])

sinlib-0.2.2/tests/test_visualize.py

@@ -0,0 +1,15 @@
+import os
+import matplotlib.font_manager as fm
+from sinlib import setup_matplotlib
+from sinlib.utils import setup_matplotlib as utils_setup_matplotlib
+
+
+def test_setup_matplotlib_import() -> None:
+    assert setup_matplotlib is utils_setup_matplotlib
+
+
+def test_setup_matplotlib_returns_font_properties() -> None:
+    fp = setup_matplotlib()
+    assert isinstance(fp, fm.FontProperties)
+    assert fp.get_name() == "Noto Sans Sinhala"
+    assert os.path.exists(fp.get_file())

sinlib-0.2.2/website/src/content/docs/api/visualization.mdx

@@ -0,0 +1,49 @@
+---
+title: Visualization
+description: API reference for sinlib's Matplotlib Sinhala text visualization utilities.
+tableOfContents: false
+---
+
+Utility function for setting up zero-configuration Sinhala text rendering in Matplotlib.
+
+<div class="api-two-column">
+<div class="api-desc-col">
+
+<span class="api-tag tag-method">Function</span>
+
+## setup\_matplotlib
+
+Downloads Noto Sans Sinhala from the Hugging Face Hub (if not already cached) and returns a Matplotlib `FontProperties` object configured for Sinhala + English rendering.
+
+<div class="api-signature">
+{`setup_matplotlib()`}
+</div>
+
+### Returns
+- `matplotlib.font_manager.FontProperties`: Font object to pass to Matplotlib text rendering calls via `fontproperties=fp`.
+
+</div>
+
+<div class="api-code-col">
+
+### Code Example
+
+```python
+import matplotlib.pyplot as plt
+from sinlib import setup_matplotlib
+
+fp = setup_matplotlib()
+
+fig, ax = plt.subplots()
+ax.text(
+    0.5, 0.5,
+    "Target Token: 'සිංහල'",
+    fontsize=20,
+    ha="center",
+    fontproperties=fp
+)
+plt.savefig("output.png")
+```
+
+</div>
+</div>

sinlib-0.2.2/website/src/content/docs/examples/visualization.md

@@ -0,0 +1,33 @@
+# Matplotlib Sinhala Text Visualization
+
+`sinlib` provides a zero-configuration utility `setup_matplotlib()` to resolve and configure Sinhala text rendering in Matplotlib.
+
+## Quickstart
+
+```python
+import matplotlib.pyplot as plt
+from sinlib import setup_matplotlib
+
+# Fetch and prepare font properties (Noto Sans Sinhala)
+fp = setup_matplotlib()
+
+# Create a plot with mixed English and Sinhala text
+fig, ax = plt.subplots(figsize=(8, 3))
+ax.set_title("OCR Model Attention Visualizer", fontsize=14, fontproperties=fp)
+ax.text(
+    0.5, 0.5,
+    "Predicted Text: 'ඇසළ පුර පසලොස්වක පොහෝ දිනය අදයි'",
+    fontsize=16,
+    ha="center",
+    va="center",
+    fontproperties=fp
+)
+ax.axis("off")
+plt.savefig("attention_vis.png", dpi=200, bbox_inches="tight")
+```
+
+## Features
+
+- **Mixed-Script Support:** Noto Sans Sinhala covers both Latin (ASCII/English) and Sinhala codepoints, so strings containing both languages render cleanly in a single `plt.text()` call.
+- **Automatic Caching:** Downloads the font from Hugging Face Hub on first run and caches it locally via `huggingface_hub`.
+- **No Global Mutation:** Returns a clean `FontProperties` instance instead of mutating Matplotlib's global `rcParams`.