rhoknp 0.4.1

Yet another Python binding for Juman++/KNP

rhoknp: Yet another Python binding for Juman++/KNP

rhoknp is a Python binding for Juman++ and KNP.

import rhoknp

# Perform language analysis by Juman++
jumanpp = rhoknp.Jumanpp()
sentence = jumanpp.apply("電気抵抗率は電気の通しにくさを表す物性値である。")

# Save language analysis by Juman++
with open("result.jumanpp", "wt") as f:
    f.write(sentence.to_jumanpp())

# Load language analysis by Juman++
with open("result.jumanpp", "rt") as f:
    sentence = rhoknp.Sentence.from_jumanpp(f.read())

# Perform language analysis by KNP
knp = rhoknp.KNP()
sentence = knp.apply(sentence)  # or knp.apply("電気抵抗率は...")

# Save language analysis by KNP
with open("result.knp", "wt") as f:
    f.write(sentence.to_knp())

# Load language analysis by KNP
with open("result.knp", "rt") as f:
    sentence = rhoknp.Sentence.from_knp(f.read())

Requirements

• Python 3.8+
• Juman++ v2.0.0-rc3+
• KNP 5.0+

Installation

pip install rhoknp

Documentation

https://rhoknp.readthedocs.io/en/latest/

Quick tour

rhoknp provides APIs to perform language analysis by Juman++ and KNP.

# Perform language analysis by Juman++
jumanpp = rhoknp.Jumanpp()
sentence = jumanpp.apply("電気抵抗率は電気の通しにくさを表す物性値である。")

# Perform language analysis by KNP
knp = rhoknp.KNP()
sentence = knp.apply(sentence)  # or knp.apply("電気抵抗率は...")

Sentence objects can be saved in the Juman/KNP format

# Save language analysis by Juman++
with open("result.jumanpp", "wt") as f:
    f.write(sentence.to_jumanpp())

# Save language analysis by KNP
with open("result.knp", "wt") as f:
    f.write(sentence.to_knp())

and recovered from Juman/KNP-format text.

# Load language analysis by Juman++
with open("result.jumanpp", "rt") as f:
    sentence = rhoknp.Sentence.from_jumanpp(f.read())

# Perform language analysis by KNP
with open("result.knp", "rt") as f:
    sentence = rhoknp.Sentence.from_knp(f.read())

It is easy to access the linguistic units that make up a sentence.

for clause in sentence.clauses:
    ...
for phrase in sentence.phrases:  # a.k.a. bunsetsu
    ...
for base_phrase in sentence.base_phrases:  # a.k.a. kihon-ku
    ...
for morpheme in sentence.morphemes:
    ...

rhoknp also provides APIs for document-level language analysis.

document = rhoknp.Document.from_raw_text(
    "電気抵抗率は電気の通しにくさを表す物性値である。単に抵抗率とも呼ばれる。"
)
# If you know sentence boundaries, you can use `Document.from_sentences` instead.
document = rhoknp.Document.from_sentences(
    [
        "電気抵抗率は電気の通しにくさを表す物性値である。",
        "単に抵抗率とも呼ばれる。",
    ]
)

Document objects can be handled in almost the same way as Sentence objects.

# Perform language analysis by Juman++/KNP
document = jumanpp.apply_to_document(document)
document = knp.apply_to_document(document)

# Save language analysis by Juman++/KNP
with open("result.jumanpp", "wt") as f:
    f.write(document.to_jumanpp())
with open("result.knp", "wt") as f:
    f.write(document.to_knp())

# Load language analysis by Juman++/KNP
with open("result.jumanpp", "rt") as f:
    document = rhoknp.Document.from_jumanpp(f.read())
with open("result.knp", "rt") as f:
    document = rhoknp.Document.from_knp(f.read())

# Access language units in the document
for sentence in document.sentences:
    ...
for clause in document.clauses:
    ...
for phrase in document.phrases:
    ...
for base_phrase in document.base_phrases:
    ...
for morpheme in document.morphemes:
    ...

For more information, explore the examples and documentation.

Main differences from pyknp

• Support document-level language analysis: rhoknp can load and instantiate the result of document-level language analysis: i.e., cohesion analysis and discourse relation analysis.
• Strictly type-aware: rhoknp is thoroughly annotated with type annotations. Efficient development is possible with the help of an IDE.
• Extensive test suite: rhoknp is tested with an extensive test suite. See the code coverage at Codecov.
• Support Python3.8+ only

Reference

• KNP FORMAT
• KNP - KUROHASHI-CHU-MURAWAKI LAB