Skip to main content

NLP as a Service

Project description

Natural Language Processing API

API Key Build Coverage Status Version Downloads Discord

One AI is a NLP as a service platform. Our API enables language comprehension in context, transforming texts from any source into structured data to use in code.

This SDK provides safe and convenient access to One AI's API from a Python environment.

Documentation

See the documentation

Getting started

Requirements

Python 3.7+ (PyPy supported)

Installation

pip install oneai

Authentication

You will need a valid API key for all requests. Register and create a key for your project in the Studio.

Example

import oneai

oneai.api_key = '<YOUR-API-KEY>'
pipeline = oneai.Pipeline(steps=[
    oneai.skills.Names(),
    oneai.skills.Summarize(min_length=20),
    oneai.skills.Keywords()
])

my_text = 'analyze this text.'
output = pipeline.run(my_text)
print(output)

Pipeline API

Language Skills

A Language Skill is a package of trained NLP models, available via API. Skills accept text as an input in various formats, and respond with processed texts and extracted metadata.

Pipelines

Language AI pipelines allow invoking and chaining multiple Language Skills to process your input text with a single call. Pipelines are defined by listing the desired Skills.

Language Studio

The Language Studio provides a visual interface to experiment with our APIs and generate calls to use in code. In the Studio you can craft a pipeline and paste the generated code back into your repository.

Basic Example

Let's say you're interested in extracting keywords from the text.

import oneai

oneai.api_key = '<YOUR-API-KEY>'
pipeline = oneai.Pipeline(steps=[
    oneai.skills.Keywords()
])

my_text = 'analyze this text.'
output = pipeline.run(my_text)
print(output)

Multi Skills request

Let's say you're interested in extracting keywords and sentiments from the text.

import oneai

oneai.api_key = '<YOUR-API-KEY>'
pipeline = oneai.Pipeline(steps=[
    oneai.skills.Keywords(),
    oneai.skills.Sentiments()
])

my_text = 'analyze this text.'
output = pipeline.run(my_text)
print(output)

Analyzer Skills vs Generator Skills

Skills can do either text analysis, and then their output are labels and spans (labels location in the analyzed text), or they can be generator skills, in which case they transform the input text into an output text.

Here's an example for a pipeline that combines both type of skills. It will extract keywords and sentiments from the text, and then summarize it.

import oneai

oneai.api_key = '<YOUR-API-KEY>'
pipeline = oneai.Pipeline(steps=[
    oneai.skills.Keywords(),
    oneai.skills.Sentiments(),
    oneai.skills.Summarize()
])

my_text = 'analyze this text.'
output = pipeline.run(my_text)
print(output)

Order is Important

When the pipeline is invoked, it is invoked with an original text you submit. If a generator skill is ran, then all following skills will use its generated text rather then the original text. In this example, for instance, we change the order of the pipeline from the previous example, and the results will be different. Instead of extracting keywords and sentiments from the original text, keywords and sentiments will be extracted from the generated summary.

import oneai

oneai.api_key = '<YOUR-API-KEY>'
pipeline = oneai.Pipeline(steps=[
    oneai.skills.Summarize(),
    oneai.skills.Keywords(),
    oneai.skills.Sentiments()
])

my_text = 'analyze this text.'
output = pipeline.run(my_text)
print(output)

Configuring Skills

Many skills are configurable as you can find out in the docs. Let's use the exact same example, this time however, we'll limit the summary length to 50 words.

import oneai

oneai.api_key = '<YOUR-API-KEY>'
pipeline = oneai.Pipeline(steps=[
    oneai.skills.Summarize(max_length=50),
    oneai.skills.Keywords(),
    oneai.skills.Sentiments()
])

my_text = 'analyze this text.'
output = pipeline.run(my_text)
print(output)

Output

The structure of the output is dynamic, and corresponds to the Skills used and their order in the pipeline. Each output object contains the input text (which can be the original input or text produced by generator Skills), and a list of labels detected by analyzer Skills, that contain the extracted data. For example:

pipeline = oneai.Pipeline(steps=[
    oneai.skills.Sentiments(),
    oneai.skills.Summarize(max_length=50),
    oneai.skills.Keywords(),
])

my_text = '''Could a voice control microwave be the new norm? The price is unbeatable for a name brand product, an official Amazon brand, so you can trust it at least. Secondly, despite the very low price, if you don't want to use the voice control, you can still use it as a regular microwave.'''
output = pipeline.run(my_text)

will generate the following:

oneai.Output(
    text="Could a voice control microwave be the ...",
    sentiments=[ # list of detected sentiments
        oneai.Label(
            type='sentiment',
            output_spans=[ # where the sentiment appears in the text
                Span(
                    start=49,
                    end=97,
                    section=0,
                    text='The price is unbeatable for a name brand product'
                )
            ],
            value='POS' # a positive sentiment
        ),
        ...
    ],
    summary=oneai.Output(
        text='The price is unbeatable for a name brand product, an official Amazon brand, so you can trust it at least. Despite the very low price, you can still use it as a regular microwave.',
        keywords=[ # keyword labels
            oneai.Label(type='keyword', name='price', output_spans=[Span(start=4, end=9, section=0, text='price')], value=0.253), ...
        ]
    )
)

File Uploads

Our API supports the following file extensions:

  • .txt- text content
  • .json- conversations in the One AI conversation format
  • .srt- analyze captions as conversations
  • .wav- audio files to be transcribed & analyzed
  • .jpg- detect text in pictures via OCR Upload a file by passing the the FileIO object to the pipeline
with open('./example.txt', 'r') as inputf:
    pipeline = oneai.Pipeline(steps=[...])
    output = pipeline.run(inputf)

For large audio files, use the asyncronous Pipeline.run_async

with open('./example.mp3', 'rb') as inputf:
    pipeline = oneai.Pipeline(steps=[oneai.skills.Transcribe(), ...])
    output = await pipeline.run(inputf)

Support

Feel free to submit issues in this repo, contact us at devrel@oneai.com, or chat with us on Discord

Project details


Release history Release notifications | RSS feed

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

oneai-0.9.89.tar.gz (33.7 kB view details)

Uploaded Source

Built Distribution

oneai-0.9.89-py3-none-any.whl (32.1 kB view details)

Uploaded Python 3

File details

Details for the file oneai-0.9.89.tar.gz.

File metadata

  • Download URL: oneai-0.9.89.tar.gz
  • Upload date:
  • Size: 33.7 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.1 CPython/3.9.6

File hashes

Hashes for oneai-0.9.89.tar.gz
Algorithm Hash digest
SHA256 7dc4b7d003b01dbf1b4e981b311a71686c244cfa0c0950ec2da8003ea542addd
MD5 2959b5bc4d47450ba41321dce69b6640
BLAKE2b-256 48eebc1e2376526ac25a4c18a7968e5bad8a1deae9ea6bc9e7305b938437425d

See more details on using hashes here.

File details

Details for the file oneai-0.9.89-py3-none-any.whl.

File metadata

  • Download URL: oneai-0.9.89-py3-none-any.whl
  • Upload date:
  • Size: 32.1 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.1 CPython/3.9.6

File hashes

Hashes for oneai-0.9.89-py3-none-any.whl
Algorithm Hash digest
SHA256 761035b5f8159a92352b0c9b831dac7bf537b2004ce922cc1e733198fd29e463
MD5 b9ca6c42c2b5d908a89138cd4e6dde75
BLAKE2b-256 3f18ddef35019434d6cbcff7a73e7708f510ccd2e9602a667b12cf7597285e12

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page