Saa

Making Time Speak! 🎙️ Translating time into colloquial spoken expressions

Saa allows you to effortlessly translate time into colloquial spoken expressions. The word saa means hour|time in Swahili, and this package aims to humanify time expression across languages. It is build using pure Python standard libraries.

from datetime import time
from saa import Clock

clock = Clock("en")
clock("11:15") # 'quarter past eleven'

ur = Clock("da")
t = time(hour=7, minute=30)
ur(t) # 'halvotte'

muda = Clock("sw")
muda("7:29") # 'saa moja na dakika ishirini na tisa asubuhi'

clock = Clock("zh")
clock("11:34") # '十一点三十四'

clock = Clock("ru")
clock("12:17") # 'двенадцать семнадцать'

Features

• Convert time into spoken expressions in various languages.
• Easy-to-use API with a simple and intuitive design.
• Pure Python implementation with no external dependencies.
• Extensible architecture for adding support for additional languages using the plugin design pattern.
• Compatible with Python 3.6 and higher.

Installation

You can install Saa using pip:

pip install -U saa

Quick Start

Here's a simple example to get you started:

from saa import Clock

# Create a Clock instance with the desired language (e.g., "en" for English)
clock = Clock("en")

# Translate time into a human-friendly spoken expression
# supports also datetime and time. .e.g. time(hour=11, minute=45)
spoken_time = clock("11:45")

print(spoken_time)
# Output: "quarter to twelve"

Using Saa with Swarm 🦜🔗

from datetime import datetime
from typing import Literal

from openai import OpenAI
from saa import Clock
from swarm import Agent, Swarm

ollama_client = OpenAI(base_url="http://localhost:11434/v1", api_key="NotNeeded",)

client = Swarm(client=ollama_client)


def get_spoken_time(language: Literal["en", "da", "sw",]) -> str:
    """
    Get spoken time:
       language: iso code for a language
    """
    clock = Clock(language)
    now = datetime.now()
    results = clock(now)
    print(results)
    return results


def transfer_to_ur(message) -> Agent:
    print(f"The Message: {message}")
    _ = message
    return ur


# list of open source model supporting tools https://ollama.com/search?c=tools

preben = Agent(
    name="Preben",
    instructions="You are Preben. Function operator",
    functions=[transfer_to_ur],  # type: ignore
    model="smollm2",
)

ur = Agent(
    name="Ur",
    instructions="Use get_spoken_time to answer what time it is a given language.",
    functions=[get_spoken_time],  # type: ignore
    model="qwen2.5",
)

response = client.run(
    agent=preben,
    messages=[{"role": "user", "content": "Ask what time it is in Danish?"}],
)

print(response.messages[-1]["content"])

---

Adding New Language 💾

Using Kiswahili as an example of how to add a new language

1. Create a folder under saa/luga directory, using the ISO 639-1 language code, with dunder init python file.

mkdir saa/luga/sw && touch saa/luga/sw/__init__.py

2. Contents of __init__.py must have the following pattern

from dataclasses import dataclass
from saa.core.language import Luga


@dataclass(init=False, eq=False, repr=False, frozen=False)
class LanguageName(Luga):
   ...

class Language(LanguageName):
   pass

So for Swahili the skeleton of saa/luga/sw/__init__.py would be:

...

@dataclass(init=False, eq=False, repr=False, frozen=False)
class Swahili(Luga):
    ...

class Language(Swahili):
    pass
...

Since we are implementing Luga, our tasks now are to implement both the properties (time, number_connector, connect_format) and static methods (time_logic, post_logic).

In Swahili class, time is expressed in terms of hour first and then minutes. time_indicator is a placeholder for whatever changing logic, e.g. minute or minutes. For Swahili the logic is what part of the day it is, morning, noon, evenning or night.

time = {
    "past": "saa {hour} na dakika {minute} time_indicator",
    "to": "saa {hour} kasoro dakika {minute} time_indicator",
    0: "saa {hour} time_indicator",
    15: "saa {hour} na robo time_indicator",
    45: "saa {hour} kasorobo time_indicator",
    30: "saa {hour} na nusu time_indicator",
}

The numbers connector is Swahili is na, and the connection format is "{tens_digits @ index 0} {[number_connector] @ index 1} {ones_digits @ index 2}"

number_connector = "na"
connect_format = "{0} {1} {2}"

Given the implementations of Numbers converter, will include 11-19 even though we could deduced them as in 20-50s. The numbers are as following:

numbers = {
    0: "sifuri",
    1: "moja",
    2: "mbili",
    3: "tatu",
    4: "nne",
    5: "tano",
    6: "sita",
    7: "saba",
    8: "nane",
    9: "tisa",
    10: "kumi",
    11: "kumi na moja",
    12: "kumi na mbili",
    13: "kumi na tatu",
    14: "kumi na nne",
    15: "kumi na tano",
    16: "kumi na sita",
    17: "kumi na saba",
    18: "kumi na nane",
    19: "kumi na tisa",
    20: "ishirini",
    30: "thelathini",
    40: "arobaini",
    50: "hamsini",
}

The major task is on time logic. In Swahili, 7 AM is the first hour in the morning (asubuhi), while 7 PM is the first hour in the evenning (jioni). 6 AM is the 12th hour in the morning ( asubuhi), while 6 PM is the 12th hour in the evenning(jioni).

"""
 0 - 11 asubuhi 
 12 - 15 mchana 
 16 - 19 jioni
 20 - 23 usiku 
"""

day_divisions = {
        0: "asubuhi",
        1: "asubuhi",
        2: "asubuhi",
        3: "asubuhi",
        4: "asubuhi",
        5: "asubuhi",
        6: "asubuhi",
        7: "asubuhi",
        8: "asubuhi",
        9: "asubuhi",
        10: "asubuhi",
        11: "asubuhi",
        12: "mchana",
        13: "mchana",
        14: "mchana",
        15: "mchana",
        16: "jioni",
        17: "jioni",
        18: "jioni",
        19: "jioni",
        20: "usiku",
        21: "usiku",
        22: "usiku",
        23: "usiku",
    }

    @staticmethod
    def post_logic(text: str) -> str:
        return text

Time to write tests. Duplicate English or Danish test_number and test_clock folders, and replace them with new language. See Swahili tests.

---

Supported Languages

Saa currently supports the following languages:

• English (en)
• Danish (da)
• Swahili (sw)
• Chinese (zh)
• Russian (ru)
• German (de) # review needed

Coming ...

• French (fr)
• Spanish (es)
• Italian (it)

Extending Language Support

One of the key strengths of Saa is its extensible architecture, allowing you to easily add support for additional languages. To add a new language, follow these steps:

1. Create a new directory under the saa/luga directory, using the ISO 639-1 language code as the filename (e.g., fr for French) and create a Python __init__.py.
2. Implement the necessary functions in the new file to translate time into spoken expressions of the target language.
3. Test the new language integration thoroughly to ensure accurate and reliable translations.
4. Consider submitting a pull request to contribute your new language support to the main Saa repository.

We welcome contributions from the community to expand language support and make Saa even more versatile!

Contributing

If you'd like to contribute to Saa, please follow the guidelines outlined in the CONTRIBUTING.md file. We appreciate your help in making this package better.

License

Saa is released under the MIT License. Feel free to use, modify, and distribute this package as per the terms of the license.

Acknowledgments

We would like to express our future gratitude to the developers of French, Spanish, German, and Italian language plugins for their valuable contributions to the Saa package. 🤣

Contact

For any questions, suggestions, or feedback, please reach out to our team at praysonpigmail.com.

Let Saa simplify time for you and enhance the way you communicate it across languages!