AI_SUMMARIZE menghasilkan ringkasan dengan memanggil Large Language Model (LLM) melalui operator Flink. Fitur ini memerlukan Ververica Runtime (VVR) 11.4 atau versi yang lebih baru dan dapat mengalami backpressure atau timeout akibat rate limiting pada platform model. - Realtime Compute for Apache Flink

AI_SUMMARIZE adalah fungsi bernilai tabel yang memanggil large language model (LLM) untuk menghasilkan ringkasan teks dari string input.

Batasan

Memerlukan Ververica Runtime (VVR) 11.4 atau versi yang lebih baru.
Throughput dibatasi oleh rate limit platform model yang mendasarinya. Ketika batas tersebut tercapai, pekerjaan Flink mengalami backpressure dan AI_SUMMARIZE menjadi bottleneck. Dalam kasus parah, hal ini dapat memicu error timeout dan restart pekerjaan.

Sintaksis

AI_SUMMARIZE(
  MODEL => MODEL <model_name>,
  INPUT => <input_column>,
  MAX_LENGTH => <max_length>
)

Mendukung argumen posisional maupun bernama.

Parameter

Parameter	Tipe data	Deskripsi
`MODEL <model_name>`	MODEL	Nama layanan model yang telah didaftarkan. Model tersebut harus mengembalikan output bertipe `VARIANT`. Lihat Model Settings untuk mendaftarkan layanan model.
`<input_column>`	STRING	Kolom yang isinya diringkas oleh model.
`<max_length>`	INTEGER	Panjang maksimum output model. Harus berupa nilai konstan.

Output

Kolom	Tipe data	Deskripsi
`summary`	STRING	Ringkasan yang dihasilkan.

Contoh

Contoh lengkap dengan data tabel

Contoh berikut membuat model Qwen-Plus, memuat data uji ke dalam tampilan sementara, dan memanggil AI_SUMMARIZE pada kolom tabel menggunakan argumen posisional maupun bernama.

Data uji

id	description
1	What is Flink? Apache Flink is an open source distributed stream processing framework for stateful computation over real-time data streams and batch data. In simple terms: Flink is a compute engine for processing real-time data. It handles continuous data streams such as website clicks, Internet of Things sensor data, and stock trades. It provides low latency, high throughput, and exactly-once semantics. It supports both stream processing and batch processing.

description

What is Flink? Apache Flink is an open source distributed stream processing framework for stateful computation over real-time data streams and batch data. In simple terms: Flink is a compute engine for processing real-time data. It handles continuous data streams such as website clicks, Internet of Things sensor data, and stock trades. It provides low latency, high throughput, and exactly-once semantics. It supports both stream processing and batch processing.

SQL

CREATE TEMPORARY MODEL general_model
INPUT (`input` STRING)
OUTPUT (`content` VARIANT)
WITH (
    'provider' = 'openai-compat',
    'task' = 'chat/completions',
    'model' = 'qwen-plus'
);

CREATE TEMPORARY VIEW infos(id, description)
AS VALUES (1, '
What is Flink?
Apache Flink is an open source distributed stream processing framework for stateful computation over real-time data streams and batch data.
In simple terms:
Flink is a compute engine for processing real-time data.
It handles continuous data streams such as website clicks, Internet of Things sensor data, and stock trades.
It provides low latency, high throughput, and exactly-once semantics.
It supports both stream processing and batch processing.
');

-- Gaya argumen posisional
SELECT id, summary
FROM infos, LATERAL TABLE(
  AI_SUMMARIZE(MODEL general_model, description, 10));

-- Gaya argumen bernama
SELECT id, summary
FROM infos, LATERAL TABLE(
  AI_SUMMARIZE(
    MODEL => MODEL general_model,
    INPUT => description,
    MAX_LENGTH => 10));

Output

id	summary
1	Real-time stream processing engine