<p align="center">
<img src="https://raw.githubusercontent.com/MahdiAlikhah/rishe/main/docs/banner.jpg" alt="Rishe: fast Persian NLP and morphology, built in Rust" width="100%">
</p>
# Rishe (ریشه)
**Fast Persian NLP and morphology, built in Rust.**
Rishe (*root* in Persian) turns raw Persian text into clean, structured data in
one call: normalized text, tokens, lemmas, sentences, ezafe markers, and a
formal rewrite of colloquial writing. It also ships the morphology utilities
that usually live in a separate library: stemming, verb conjugation,
inflection, token splitting, spell correction, and stopword removal.
The core is Rust. You get it as a `pip install` with prebuilt wheels, no model
downloads, no runtime dependencies, and the same engine as a Rust crate.
```python
from rishe import Pipeline
pipe = Pipeline()
doc = pipe("اون داره تو خونه کتاب میخونه، چیه؟")
doc.tokens # ['آن', 'دارد', 'تو', 'خانه', 'کتاب', 'میخواند', '،', 'چیست', '؟']
doc.lemmas # ['آن', 'داشت', 'تو', 'خانه', 'کتاب', 'خواند', '،', 'چیست', '؟']
doc.formal_text # 'آن دارد تو خانه کتاب میخواند، چیست؟'
```
## Why Rishe
- **Normalization about 80× faster** than Hazm and Shekar: 1.2 µs per sentence
against 96 µs and 104 µs.
- **Best tokenization and lemmatization in this comparison.** Token F1 of
1.0000 on UD-Seraji, 0.9989 on UD-PerDT and 0.9028 on PEYMA; lemma accuracy
of 0.9209 and 0.9316, ahead of Hazm (0.9048 / 0.9009) and Shekar (0.81 / 0.82).
- **Morphology in the same core.** Stemming, conjugation, inflection, token
splitting, spell correction and stopwords — no second library, no extra
runtime. On the UD lemma proxy, stemming is 89.2% / 89.5% exact against Hazm
(66.4% / 64.4%) and Shekar (76.1% / 72.0%).
- **Spell correction that is fast enough to stay in the loop.** 80.8% top-1 and
96.2% top-5 on synthetic 1-edit typos, at 0.16 ms per word; Shekar's
statistical checker is 60.8% / 77.6% at 94 ms per word on the same set.
- **Millions of tokens per second.** 1.0M tokens/s on one thread; `pipe()`
runs batches in parallel with the GIL released and reaches 6.5M tokens/s on
12 threads.
- **Everything in one call.** Normalization, informal → formal conversion,
tokenization, lemmatization, sentence splitting and ezafe detection from one
`Pipeline` object, in about 16 µs per sentence.
- **Informal Persian handled.** Colloquial forms (`اون`, `خونه`, `داره`,
`میخوام`, `چیه`) are rewritten to their formal spelling, and fused clitics
such as `بهش` and `ازش` are split into `به` + `او` and `از` + `او`.
- **Nothing to set up.** Compiled-in lexicons (190k nouns, 1.2k verb stems, word
frequencies, stopwords), no model downloads, no Python dependencies, no
network access.
- **Wheels for every major platform.** Linux (glibc and musl, x86_64 and
aarch64), macOS (Intel and Apple Silicon) and Windows, for Python 3.9+
(abi3). The same core is available as a Rust crate.
- **Thread-safe.** One `Pipeline` can be shared across threads.
## Benchmarks
Measured against [Hazm](https://github.com/roshan-research/hazm) and
[Shekar](https://github.com/amirivojdan/shekar) on the full UD Persian-Seraji,
UD Persian-PerDT and PEYMA test sets. Every library is built once and reused.
**Core speed** (per sentence, single thread)
| Normalize | **1.2 µs** | 96 µs | 104 µs |
| Tokenize | **4.1 µs** | 5.7 µs | 4.4 µs |
| Full pipeline | **16.3 µs** | — | — |
**Core accuracy**
| Token F1 | UD-Seraji | **1.0000** | Shekar 0.9992, Hazm 0.9810 |
| Token F1 | UD-PerDT | **0.9989** | Shekar 0.9976, Hazm 0.9609 |
| Token F1 | PEYMA | **0.9028** | Shekar 0.9018, Hazm 0.8901 |
| Lemma accuracy | UD-Seraji | **0.9209** | Hazm 0.9048, Shekar 0.8058 |
| Lemma accuracy | UD-PerDT | **0.9316** | Hazm 0.9009, Shekar 0.8170 |
**New feature highlights**
| Stemming | exact UD-lemma proxy (Seraji / PerDT) | **89.2% / 89.5%** | 66.4% / 64.4% | 76.1% / 72.0% |
| Conjugation | hand-written spot-check (29 forms) | **29/29** | 28/29 | 29/29 |
| Conjugation | ms/verb | **0.010** | 0.152 | 0.080 |
| Inflection | comparative / superlative / plural | **100%** | N/A | **100%** |
| Token splitting | ms/token | **0.0006** | 12.05 | N/A |
| Spell correction | top-1 / top-5 on 1-edit typos | **80.8% / 96.2%** | N/A | 60.8% / 77.6% |
| Stopwords | list size | 389 | 389 | 301 |
The normalizer operators are 1.5–6× faster than Shekar's per operator (the
exception is `ya`, 0.7×) with exact output agreement of 100% on 8 of 11
operators and 99.04% on verbal spacing. `normalize_stages` is opt-in; the default
`normalize` path is unchanged.
The full core tables, batch scaling and method are in
[docs/BENCHMARKS.md](https://github.com/MahdiAlikhah/rishe/blob/main/docs/BENCHMARKS.md).
The feature tables, proxy metrics and caveats are in
[docs/FEATURE_BENCHMARKS.md](https://github.com/MahdiAlikhah/rishe/blob/main/docs/FEATURE_BENCHMARKS.md).
Reproduce:
```bash
python scripts/benchmark.py # core token/lemma/speed
python scripts/benchmark_features.py # stemming, conjugation, splitting, spelling
python scripts/benchmark_normalizer.py # 11 normalizer operators vs Shekar/Hazm
python scripts/benchmark_maskers.py # 12 maskers vs Shekar
```
## Install
```bash
pip install rishe # Python >= 3.9
```
or, from Rust (`Cargo.toml`): `rishe = "0.2"`.
No Python dependencies. No model downloads. No network access at runtime.
## Usage
```python
from rishe import Pipeline
pipe = Pipeline() # unify digits, strip diacritics, fix ZWNJ
doc = pipe("کتابهای خوب را به دوستانم میدهم.")
doc.normalized_text, doc.formal_text, doc.sentences
doc.tokens, doc.lemmas, doc.ezafe_markers
docs = pipe.pipe(texts) # parallel batch -> list[ProcessedDoc]
pipe.normalize("مي كنم 2026") # 'میکنم ۲۰۲۶'
pipe.tokenize("سلام، دنیا!") # raw tokenization, no normalization
pipe.word_tokenize("سلام، دنیا!") # normalize + tokenize
pipe.sent_tokenize(text) # sentence splitting only
pipe.lemmatize("میروم") # 'رفت' (past-stem lemmas, UD style)
pipe.lemma_full("میروم") # ('رفت', 'رو') (past + present stem)
```
`pipe(text)` also rewrites colloquial words to formal ones, so its `tokens` can
differ from the surface text; `word_tokenize` tokenizes without rewriting.
`tokenize` is the fast raw path used for the tokenization benchmarks.
### Morphology and utilities
```python
pipe.stem("کتابها") # 'کتاب'
pipe.conjugate("رفت")["present"][:3] # ['میروم', 'میروی', 'میرود']
pipe.comparative("بزرگ") # 'بزرگتر'
pipe.superlative("بزرگ") # 'بزرگترین'
pipe.plural("کتاب") # 'کتابها'
pipe.possessive("خانه", 0) # 'خانهام'
pipe.split_token("بهش") # ['به', 'او']
pipe.spell_suggest("سسلام") # ['سلام', 'اسلام', ...]
pipe.spell_correct("سسلام بر ششما") # 'سلام بر شما'
pipe.tokenize_spans("سلام دنیا") # [('سلام', 0, 4), ('دنیا', 5, 9)] (char offsets)
pipe.remove_stopwords(tokens) # drop Persian stopwords
pipe.process_stages(text, ["normalize", "sentences"]) # choose pipeline stages
```
### Normalizer operators
`normalize_stages` applies the P1 normalizer operators in the order you list them:
`alphabet`, `arabic_unicode`, `ya`, `ya_joda`, `digits`, `punctuation`,
`repeated_letters`, `orthographic_spacing`, `punctuation_spacing`, `word_spacing`,
`verbal_spacing`, and `spacing` (all four spacing operators).
```python
pipe.normalize_stages("﷽ 1𝟮3٤٥⓺", ["arabic_unicode", "digits"])
# 'بسم الله الرحمن الرحیم ۱۲۳۴۵۶'
pipe.normalize_stages("سلام(دنیا) ! می روم", ["punctuation_spacing", "verbal_spacing"])
# 'سلام (دنیا)! میروم'
```
`word_spacing` and `verbal_spacing` validate against the compiled lexicons, so they
only join forms the engine knows (`کتاب ها` → `کتابها`, `رفته ام` → `رفتهام`).
Unknown stage names raise `ValueError`.
### Masking
`mask_stages` removes or masks URLs, emails, mentions, hashtags, emoji, HTML
tags, non-Persian letters, stopwords, offensive words, digits, diacritics and
punctuation. Stage names are `url`, `email`, `mention`, `hashtag`, `emoji`,
`html`, `non_persian`, `stopwords`, `offensive`, `digits`, `diacritics` and
`punctuation`.
```python
pipe.mask_stages("سلام @user، به https://rishe.dev یا info@rishe.dev بنویس.",
["url", "email", "mention"])
# 'سلام به <URL> یا <EMAIL> بنویس.'
pipe.mask_stages("این کتاب است", ["stopwords"], mask_token="[MASK]")
# '[MASK] کتاب [MASK]'
```
Defaults are `<URL>` for URLs, `<EMAIL>` for emails, a space for mentions,
hashtags and HTML tags, and removal for everything else. Pass `mask_token` to
override all stages. Masking keeps the spaces left behind; run
`normalize_stages(..., ["spacing"])` afterwards to collapse them. Unknown stage
names raise `ValueError`.
Spell correction uses a 190k-word frequency lexicon and edit-distance 1; typos
with two or more edits are left unchanged.
### Utilities
```python
pipe.number_to_words(1404) # 'یک هزار و چهارصد و چهار'
pipe.number_to_words("3.14") # 'سه ممیز چهارده'
pipe.ngrams("این یک متن", 1, 2) # ['این', 'یک', 'متن', 'این یک', 'یک متن']
pipe.persianize("فایل را با کامپیوتر باز کن") # 'پرونده را با رایانه باز کن'
pipe.persianize_suggest("فایل") # [('فایل', ['پرونده'], 0, 4)]
flatten(["الف", ["ب", "پ"]]) # 'الف', 'ب', 'پ'
```
`persianize` uses a built-in curated loanword map; `persianize_suggest` returns
the alternatives and character spans.
### Augmentation
`keyboard_noise`, `ocr_noise` and `whitespace_noise` simulate typing, OCR and
spacing errors. Pass `seed=` for reproducible output.
```python
pipe.keyboard_noise(text, seed=42)
pipe.ocr_noise(text, seed=42)
pipe.whitespace_noise(text, seed=42)
```
### Keywords and transliteration
```python
pipe.rake(text, top_n=5)
pipe.textrank(text, top_n=5)
pipe.to_tajik("خانه") # 'хона'
pipe.to_persian("хона") # 'خانه'
```
Transliteration is a rule-based script mapping; it preserves the consonant
skeleton and is not a phonetic translation.
### Tokenizer options
```python
pipe.word_tokenize_with("رفته است", join_verb_parts=True) # ['رفته_است']
pipe.word_tokenize_with("info@a.co", replace_emails=True) # ['EMAIL']
```
Options: `join_verb_parts`, `join_abbreviations`, `separate_emoji`,
`replace_links`, `replace_ids`, `replace_emails`, `replace_numbers`,
`replace_hashtags`. All are off by default, so plain `word_tokenize` is
unchanged.
### Corpora
```python
from rishe import ConlluReader, PeymaReader, PlainTextReader
for sentence in ConlluReader("data/corpora/fa_perdt-ud-test.conllu"):
print(sentence.text, sentence.tokens[0].lemma)
```
`ConlluReader` parses UD/CoNLL-U, `PeymaReader` parses the PEYMA test format,
and `PlainTextReader` reads one sentence per line. Other corpus-specific formats
can be added as needed.
### CLI
```bash
rishe normalize "مي كنم"
rishe tokenize "سلام دنیا" --json
rishe mask "info@a.co" --stages email
printf '%s\n' '{"command":"tokenize","text":"سلام دنیا"}' | rishe serve
```
### Migrating from Hazm
```python
from rishe import Lemmatizer, Normalizer, WordTokenizer, Stemmer
normalizer = Normalizer()
normalizer.normalize("مي كنم") # 'میکنم'
lemmatizer = Lemmatizer()
lemmatizer.lemmatize("میروم") # 'رفت#رو'
WordTokenizer().tokenize("رفته است") # ['رفته_است']
```
The wrappers cover `Normalizer`, `WordTokenizer`, `SentenceTokenizer`,
`Lemmatizer`, `Stemmer` and `TokenSplitter`, plus module-level `word_tokenize`
and `sent_tokenize`. Hazm constructor arguments that point at custom lexicon
files are rejected because Rishe compiles its lexicons in.
```rust
use rishe::Pipeline;
let pipe = Pipeline::create(true, true, true);
let doc = pipe.process("کتابهای خوب را به دوستانم میدهم.");
println!("{:?}", doc.tokens);
println!("{}", pipe.stem("کتابها"));
```
## Scope
Rishe is a text-preprocessing and morphology engine: normalization,
tokenization, sentence splitting, lemmatization, stemming, verb conjugation,
inflection, token splitting, spell correction, stopword removal, ezafe
detection and informal → formal conversion.
It does **not** yet do POS tagging, chunking, NER, dependency parsing,
embeddings or text classification. Those are the next optional model packs;
the base stays zero-dependency and fast. Today, use Rishe as the preprocessing
step in front of a tagger or parser.
## Building from source
```bash
cargo test
maturin develop --release # Python extension into the current venv
python scripts/benchmark.py # needs: hazm shekar conllu
python scripts/benchmark_features.py
```
`scripts/extract_lexicons.py` regenerates `data/*.tsv` and
`data/word_freq.tsv` from Hazm's data files; `build.rs` compiles them into FSTs.
## License
Dual-licensed under [Apache-2.0](https://github.com/MahdiAlikhah/rishe/blob/main/LICENSE-APACHE)
or [MIT](https://github.com/MahdiAlikhah/rishe/blob/main/LICENSE-MIT), at your
option. The bundled lexicons, stopwords and word frequencies are derived from
[Hazm](https://github.com/roshan-research/hazm) data (MIT); see
[NOTICE.md](https://github.com/MahdiAlikhah/rishe/blob/main/NOTICE.md).