PyTeaser

PyTeaser extracts a short, extractive summary from article text or a news URL. It ranks sentences using title relevance, article keywords, sentence length, and position, then removes repeated content and returns the selected sentences in the article's original order.

Requirements and installation

PyTeaser requires Python 3.10 or newer.

python -m pip install pyteaser

Optional functionality is installed separately:

python -m pip install 'pyteaser[images]'   # Pillow-backed image extraction
python -m pip install 'pyteaser[chinese]'   # jieba Chinese word segmentation
python -m pip install 'pyteaser[soup]'      # BeautifulSoup-based fallback parser
python -m pip install 'pyteaser[all]'       # all optional features

When working from a source checkout, replace pyteaser with . (or .[all]).

Summarize text

from pyteaser import summarize

title = "A sample article"
text = "The first sentence explains the announcement. A second sentence gives context. A third sentence describes what happens next."

summary = summarize(title, text, sentence_count=2, language="en")
for sentence in summary:
    print(sentence)

title may be None; article text must be a string or UTF-8 bytes. Blank text returns an empty list. sentence_count defaults to 5, must be a non-negative integer, and may be set to 0 to request no sentences. Optional max_words sets a total token budget; candidates that do not fit are skipped.

Scoring weights default to title 1.5, keyword frequency 2.0, length 1.0, and position 1.0. Pass a ScoringWeights instance or a dictionary with those feature names to tune the heuristic.

For score inspection, summarize_detailed(...) returns records with each selected sentence's original index, text in sentence, and overall score. The normal APIs continue to return plain strings.

Command line

After installation, summarize text or a UTF-8 file:

pyteaser --title "Example article" --text "First sentence. Second sentence." --sentence-count 1
pyteaser --input-file article.txt --title "Example article" --max-words 120
pyteaser --url "https://example.com/news/article" --sentence-count 3

The original Summarize(title, text) function remains available. Both APIs return a list of sentence strings.

Summarize a URL

from pyteaser import summarize_url

summary = summarize_url("https://example.com/news/article", sentence_count=3)

URL summarization extracts the page's title and article text first. If the page declares a supported language, PyTeaser uses its stopword list; otherwise English is used. You can also explicitly pass language="es", for example.

The bundled language resources cover English, Spanish, Italian, German, Swedish, Russian, French, Arabic, and Chinese. Chinese segmentation uses jieba when installed; without it, PyTeaser falls back to character tokenization. Language support is heuristic, so review summaries for your content and audience.

The legacy SummarizeUrl(url) function remains available. URL summarization returns a list, or raises ArticleFetchError when a page cannot be fetched safely and ArticleExtractionError when no title/body can be extracted. Invalid text, language, and sentence-count inputs raise TypeError or ValueError.

URL and image safety

URL fetching accepts HTTP(S) on standard ports, applies a 10-second request timeout and a 5 MiB HTML limit, validates resolved addresses and redirect targets, and blocks private/non-routable addresses by default. Image fetching is disabled by default; enable it only when needed:

from goose import Goose

article = Goose({"enable_image_fetching": True}).extract(
    url="https://example.com/news/article"
)
print(article.top_image.src)

Image downloads are capped at 15 MiB and require the images extra. Applications processing untrusted URLs should keep private-network access disabled. For trusted local development only, Goose supports allow_private_network=True; do not enable this for user-controlled URLs.

Use article extraction responsibly and comply with the target site's terms and applicable law.

Tests

The test suite is offline; URL cases use mocked responses or a local loopback server.

python -m unittest discover -v

License

PyTeaser's original code is MIT-licensed. The bundled Goose article extractor and its included resources are Apache-2.0-licensed; see LICENSE, goose/LICENSE.txt, and NOTICE for license and attribution details.