Local Extractive Text Summarizer
Selects sentences using word-frequency scores and a first-sentence bonus. It does not run TextRank or build a sentence graph.
About this local extractive text summarizer
Selects sentences using word-frequency scores and a first-sentence bonus. It does not run TextRank or build a sentence graph Applies Unicode-compliant text transformations and parsing directly in the browser tab without data leakage.
Text processing & transformation mechanics
Selects sentences using global repeated-token frequency divided by the square root of sentence token count, with a small first-sentence bonus. The current implementation is not TextRank or TF-IDF.
Worked text manipulation example
Scenario: Choose one original sentence using the secondary sentence limit of 1.
Sample input:
Second field:
Processing: Score repeated word tokens; the first Cats sentence has the highest score after its 0.15 bonus, so return that original sentence.
Illustrative output:
Character encoding & string length boundaries
Extractive frequency heuristic only, not an abstractive model, TextRank or semantic fact check. CJK word segmentation is limited.
Text formatting and transformations in Local Extractive Text Summarizer preserve character integrity. Check edge cases such as multi-line inputs and special characters before downstream use.
In-memory text processing boundary
String operations and Unicode manipulations in Local Extractive Text Summarizer execute entirely in local browser memory without transmitting input text.
Unicode Consortium & text formatting standards
Unicode Consortium standard annexes (UAX) and character encoding specifications:
- ECMAScript Unicode letter and number property matching.
- CZOA-specific sentence frequency scoring and 0.15 first-sentence bonus; not TextRank, TF-IDF, or an external summarization standard.
Local Extractive Text Summarizer verification · Audit standards
How to use it
- Paste source text and optionally enter a sentence count in the second field.
- Score sentences by repeated-word frequency with a small first-sentence bonus.
- Read selected original sentences in source order.
Frequently asked questions
How does the local extractive summarizer choose sentences?+
It splits text at `.`, `!`, `?`, `。`, `！`, and `？`, scores each sentence from global occurrences of its letter-and-number tokens longer than one JavaScript code unit, gives the first sentence a small bonus, and returns the selected original sentences in source order.
What happens when the limit is `1` for `Cats sleep. Cats purr. Dogs bark.`?+
It returns `Cats sleep.`. In this input, the first sentence has the highest score after its small first-sentence bonus.
How does the second input control the number of sentences?+
Only an empty second field defaults to `3`. Any supplied finite number is truncated toward zero and clamped from `1` to the available sentence count. `NaN`, `abc`, and `Infinity` are rejected with a local error.
Does the summary verify facts or infer document intent?+
No. It is a local extractive frequency heuristic. It does not retrieve sources, resolve references or contradictions, preserve every qualification, or produce a guaranteed abstractive summary.