dingo-python 1.0.3

Language quality evaluation tool.

English | 简体中文

Introduction

Dingo is a data quality assessment tool that helps you automatically detect data quality issues in your datasets. Dingo provides a variety of built-in detection rules and model methods, and also supports custom detection methods. It supports commonly used NLP datasets and multimodal datasets, including pre-training datasets, fine-tuning datasets, and evaluation datasets. In addition, Dingo supports various interface usage methods, including local CLI, SDK, and RESTFul API, making it easy to integrate into various evaluation platforms, such as OpenCompass，simple-evals etc.

Architecture of Dingo

QuickStart

Install dingo.

pip install dingo

Try the following SDK demo code:

from dingo.model import Model
from dingo.io import RawInputModel
from dingo.exec import Executor

input_data = {
    "eval_models": ["sft"],
    "input_path": "tatsu-lab/alpaca", # default from huggingface
    "data_format": "plaintext",
}

raw_input = RawInputModel(**input_data)
Model.apply_config(raw_input.custom_config_path)
executor = Executor.exec_map["local"](raw_input)
result = executor.evaluate()
print(result)

you can also try CLI:

python -m dingo.run.cli --input_path tatsu-lab/alpaca -e sft --data_format plaintext

Tutorials

Config

Execute

Dingo can be run locally or on a Spark cluster.

Local Mode

In addition to the aforementioned SDK calls, you can also run data evaluation locally with CLI:

python -m dingo.run.cli

The CLI parameters are as follows.

parameter name	description
-e or --eval_models	The model used to evaluate data quality.
-i or --input_path	The path of data. It can be a file or a directory.
--output_path	The path of result data.
--data_format	The format of data. It can be json, jsonl, plaintext and list json.
--dataset	The platform for data run. It can be huggingface, local and spark.
--datasource	The source of data. It can be huggingface, local and s3.
--huggingface_split	The split of huggingface.
--column_id	The column name of id in data.
--column_prompt	The column name of prompt in data.
--column_content	The column name of content in data.
--custom_config_path	The path of custom config file.
--spark_master_url	The url of spark master.
--spark_summary_save_path	The path of summary saved when run in spark.
--s3_ak	The ak of s3.
--s3_sk	The sk of s3.
--s3_endpoint_url	The url of end point in s3.
--s3_addressing_style	The style of addressing in s3.
--s3_bucket	The bucket of s3.

More information can be obtained by running the following command: python -m dingo.run.cli --help.

Spark Mode

If the scale of data is very large you can use Spark to run the project.

Firstly, create an object from SparkExecutor, and set the actual instances of SparkSession and DataFrame.

from dingo.exec.spark import SparkExecutor

spark_exec = SparkExecutor()
spark_exec.set_spark(spark_session)
spark_exec.set_input_df(spark_data_frame)

Then, convert the data and execute the rule list.

spark_exec.convert_data(column_id=['data_id'], column_prompt=['prompt'], column_content=['content'])
spark_exec.execute(["CommonSpecialCharacter", "CommonColonEnd"])

Finally, summarize and get the result data.

spark_exec.summarize()
output_df = spark_exec.get_output_df()
summary = spark_exec.get_summary()

Evaluation Results

Summary

The summary.json file is overall information about evaluation results. Here is an example:

{
    "dataset_id": "20240816_175052",
    "input_model": "default",
    "input_path": "test/data/test_local_json.json",
    "output_path": "test/outputs/20240816_175052",
    "score": 0.0,
    "num_good": 0,
    "num_bad": 2,
    "total": 2,
    "error_type_ratio": {
        "QUALITY_INEFFECTIVENESS": 0.0,
        "QUALITY_INCOMPLETENESS": 0.0,
        "QUALITY_DISUNDERSTANDABILITY": 0.0,
        "QUALITY_DISSIMILARITY": 0.0,
        "QUALITY_DISFLUENCY": 0.0,
        "QUALITY_IRRELEVANCE": 1.0,
        "QUALITY_INSECURITY": 0.0
    },
    "error_name_ratio": {
        "QUALITY_IRRELEVANCE-CommonSpecialCharacter": 1.0
    }
}

The error_ratio field shows data quality signals in seven different aspects: EFFECTIVENESS, COMPLETENESS, UNDERSTANDABILITY, SIMILARITY, FLUENCY, RELEVANCE and SECURITY.

Detailed Results

For more detailed issues found in data items, Dingo created files in a directory named with the quality signals mentioned above. Give an example. CommonColonEnd.json in the QUALITY_SIGNAL_COMPLETENESS directory is as follows:

{"data_id": "0", "prompt": "", "content": "�I am 8 years old. ^I love apple because: fuck you", "error_type": ["QUALITY_IRRELEVANCE"], "error_name": ["QUALITY_IRRELEVANCE-CommonSpecialCharacter"], "error_reason": ["�"]}
{"data_id": "1", "prompt": "", "content": "�[I like blue best. Because blue is the color of the sky. ", "error_type": ["QUALITY_IRRELEVANCE"], "error_name": ["QUALITY_IRRELEVANCE-CommonSpecialCharacter"], "error_reason": ["�"]}

We evaluated the quality of these three datasets based on Dingo.

Dataset	Dataset Type	EFFECTIVENESS	COMPLETENESS	UNDERSTANDABILITY	SIMILARITY	FLUENCY	RELEVANCE	SECURITY
SlimPajama-627B	Pretrain	0.016860	0.000175	0.002062	0.003563	0.000302	0.003767	0
Stanford_alpaca	SFT	0.001442	0.000538	0.000481	0.000231	0	0	0
MMLU	Benchmark	0.011759	0.007349	0	0	0	0	0

Rule List

Contributing

We appreciate all contributions to Dingo. Please refer to CONTRIBUTING.md for the contributing guideline.

License

This project is released under the Apache 2.0 license.