ChatGPT 3.5 注冊(cè)已經(jīng)不需要手機(jī)了，直接郵箱認(rèn)證就可以，這可真算是好消息，壞消息是 ChatGPT 4 還是要收費(fèi)。 那么 GPT-3.5 與 GPT-4 區(qū)別有多大呢，下面簡(jiǎn)單測(cè)試一下。

以從 TDengine 訂閱數(shù)據(jù)為例，TDengine 算是不太小眾的數(shù)據(jù)庫(kù)，如果連這個(gè)都能搞定，說明GPT 還是蠻厲害的。

GPT-3.5 給出的代碼

這個(gè)我是直接從OpenAI 獲取的答案

from taospy import Client

# 連接到 TDengine 服務(wù)器
conn = Client(host='your_tdengine_server_ip', user='your_username', password='your_password', database='your_database')

# 訂閱數(shù)據(jù)回調(diào)函數(shù)
def callback(data):
    print(f'Received data: {data}')

# 訂閱數(shù)據(jù)
with conn.subscribe('your_table_name', callback):
    # 在此處可以執(zhí)行其他任務(wù)，讓訂閱保持活躍
    input('Press Enter to stop the subscription...\n')

# 關(guān)閉連接
conn.close()

GPT-4 給出的代碼

這是從DevChat 簡(jiǎn)介獲取的答案

import taos
import time

# 創(chuàng)建訂閱對(duì)象
def create_subscribe(conn):
    sql = 'select * from test.meters interval(1s)'
    print("sql: ", sql)
    sub = conn.subscribe(sql, restart=False)
    print("sub name:", sub.name())
    print("create sub res:", sub.result())
    return sub

# 獲取訂閱對(duì)象的值
def get_subscribe_value(sub):
    while True:
        print("blocked get result.")
        sub.consume()
        print("result: ", sub.result.current)
        time.sleep(1)

# 刪除訂閱對(duì)象
def unsub(sub):
    print("unsub")
    sub.unsubscribe()

try:
    conn = taos.connect(host="TDengine的地址", user="root", password="taosdata", database="test", charset="UTF8")
    sub = create_subscribe(conn)
    time.sleep(2)
    get_subscribe_value(sub)
    time.sleep(2)
    unsub(sub)
    print("End of program!")
except Exception as e:
    print(e)

官方示例代碼

地址如下： python/#數(shù)據(jù)訂閱

from taos.tmq import Consumer
import taos


def init_tmq_env(db, topic):
    conn = taos.connect()
    conn.execute("drop topic if exists {}".format(topic))
    conn.execute("drop database if exists {}".format(db))
    conn.execute("create database if not exists {} wal_retention_period 3600".format(db))
    conn.select_db(db)
    conn.execute(
        "create stable if not exists stb1 (ts timestamp, c1 int, c2 float, c3 varchar(16)) tags(t1 int, t3 varchar(16))")
    conn.execute("create table if not exists tb1 using stb1 tags(1, 't1')")
    conn.execute("create table if not exists tb2 using stb1 tags(2, 't2')")
    conn.execute("create table if not exists tb3 using stb1 tags(3, 't3')")
    conn.execute("create topic if not exists {} as select ts, c1, c2, c3 from stb1".format(topic))
    conn.execute("insert into tb1 values (now, 1, 1.0, 'tmq test')")
    conn.execute("insert into tb2 values (now, 2, 2.0, 'tmq test')")
    conn.execute("insert into tb3 values (now, 3, 3.0, 'tmq test')")


def cleanup(db, topic):
    conn = taos.connect()
    conn.execute("drop topic if exists {}".format(topic))
    conn.execute("drop database if exists {}".format(db))


if __name__ == '__main__':
    init_tmq_env("tmq_test", "tmq_test_topic")  # init env
    consumer = Consumer(
        {
            "group.id": "tg2",
            "td.connect.user": "root",
            "td.connect.pass": "taosdata",
            "enable.auto.commit": "true",
        }
    )
    consumer.subscribe(["tmq_test_topic"])

    try:
        while True:
            res = consumer.poll(1)
            if not res:
                break
            err = res.error()
            if err is not None:
                raise err
            val = res.value()

            for block in val:
                print(block.fetchall())
    finally:
        consumer.unsubscribe()
        consumer.close()
        cleanup("tmq_test", "tmq_test_topic")

區(qū)別有多大呢？ 用圖片說話
文章來源地址http://www.zghlxwxcb.cn/news/detail-771203.html

到了這里，關(guān)于[AI]ChatGPT4 與 ChatGPT3.5 區(qū)別有多大的文章就介紹完了。如果您還想了解更多內(nèi)容，請(qǐng)?jiān)谟疑辖撬阉鱐OY模板網(wǎng)以前的文章或繼續(xù)瀏覽下面的相關(guān)文章，希望大家以后多多支持TOY模板網(wǎng)！