您可以尝试提取div的所有子项，并执行正则表达式测试，以查看它是h2还是h3，然后测试文本是否包含"Diet"或"Characteristics"，如果通过，则收集所有为<p>的同级。

def parse(self, response):
    collect = False
    contents = []
    for selector in response.xpath("//div[@class='entry-content']/*"):
        val = selector.xpath("./text()").get()
        if collect and selector.re('<p'):
            contents.append(val)
            continue
        if val and selector.re(r'<h[23]'):
            if "Characteristics" in val or "Diet" in val:
                collect = True
        else:
            collect = False
    yield {"contents" : contents}