EasyScrapper mapping

mail@pastecode.io avatar
unknown
json
a month ago
4.1 kB
4
Indexable
Never
{
	"title":
	{
		"primary":
		{
			"conditions":
			{
				"selector": ".product-info-right h1[itemprop=\"name\"]",
				"type": "text"
			},
			"operations":
			{
				"default": "No title",
				"remove": ["studio Lounge ", "Lounge Tee Black"]
			}
		},
		"secondary":
		{
			"conditions":
			{
				"selector": "._product-details-title ._h2",
				"type": "text",
				"operations":
				{
					"remove": "\u043b\u0432."
				}
			}
		},
		"operations":
		{
			"default": "No title"
		}
	},
	"description":
	{
		"primary":
		{
			"conditions":
			{
				"selector": "#full-description div[itemprop=\"description\"]",
				"type": "html"
			},
			"operations":
			{
				"remove_links": true,
				"remove": "Dress4Less"
			}
		}
	},
	"short_description":
	{
		"primary":
		{
			"conditions":
			{
				"selector": "ul.attributes",
				"type": "html"
			},
			"operations":
			{
				"remove_links": true,
				"remove": ["\u041c\u044a\u0436\u0435", "\u041c\u0430\u0442\u0435\u0440\u0438\u0430\u043b"]
			}
		}
	},
	"tabs":
	{
		"primary":
		{
			"conditions":
			{
				"selector_array":
				{
					".product-tabs a[data-tab-target=\"#full-description\"]": "#full-description div[itemprop=\"description\"]",
					".product-tabs a[data-tab-target=\"#delivery\"]": ".tabs-content #delivery div.text-page"
				},
				"type": "html",
				"separator_value": "li"
			},
			"operations":
			{
				"remove_links": true,
				"remove": [" \u0434\u043e\u0441\u0442\u0430\u0432\u043a\u0430 \u043f\u0440\u0438 \u043f\u043e\u0440\u044a\u0447\u043a\u0430 \u043d\u0430\u0434 90\u043b\u0432. \u043d\u0430\u043f\u0440\u0430\u0432\u0435\u043d\u0430 \u0447\u0440\u0435\u0437 \u043e\u043d\u043b\u0430\u0439\u043d \u043c\u0430\u0433\u0430\u0437\u0438\u043d\u0430! ", "<strong>\u0411\u0435\u0437\u043f\u043b\u0430\u0442\u043d\u0430 \u0438 \u0431\u044a\u0440\u0437\u0430</strong>"],
				"replace_text":
				{
					"Speedy": "Econt"
				}
			}
		}
	},
	"attributes":
	{
		"primary":
		{
			"conditions":
			{
				"selector_array":
				{
					"ul.attributes li span.label": "ul.attributes li span.value"
				},
				"separator_value": "li",
				"is_multiple": true
			},
			"operations":
			{
				"remove": [":"],
				"filter_keys": ["\u041c\u0430\u0440\u043a\u0438", "\u0422\u0435\u0433\u043b\u043e"]
			}
		}
	},
	"categories":
	{
		"template":
		{
			"name":
			{
				"primary":
				{
					"conditions":
					{
						"selector": "nav.breadcrumbs ul li a span[itemprop=\"name\"]",
						"is_multiple": true
					},
					"operations":
					{
						"remove_first": true,
						"end_with": " \u043f\u0435\u0434\u0435\u0440\u0430\u0441\u0438",
						"remove": "\u0414\u0440\u0435\u0445\u0438 "
					}
				}
			},
			"link":
			{
				"primary":
				{
					"conditions":
					{
						"selector": "nav.breadcrumbs ul li a",
						"is_multiple": true,
						"key": "href"
					},
					"operations":
					{
						"remove_first": true
					}
				}
			}
		},
		"operations":
		{
			"remove_first": true
		}
	},
	"brand":
	{
		"template":
		{
			"name":
			{
				"primary":
				{
					"conditions":
					{
						"selector": "ul.attributes li:contains(\"\u041c\u0430\u0440\u043a\u0438:\") span.value",
						"is_multiple": true
					},
					"operations":
					{
						"replace_text":
						{
							"adidas": "PUMA"
						}
					}
				}
			},
			"link":
			{
				"primary":
				{
					"conditions":
					{
						"selector": "ul.attributes li:contains(\"\u041c\u0430\u0440\u043a\u0438:\") span a",
						"is_multiple": true,
						"key": "href"
					}
				}
			},
			"image":
			{
				"primary":
				{
					"conditions":
					{
						"selector": "div.product-info-right a.brand-logo img",
						"is_multiple": true,
						"key": "src"
					}
				}
			}
		},
		"operations":
		{
			"filter_keys": ["name"]
		}
	},
	"images":
	{
		"primary":
		{
			"conditions":
			{
				"selector": ".gallery-thumbnails div.thumb img",
				"key": "srcset",
				"is_multiple": true
			}
		}
	},
	"main_image":
	{
		"primary":
		{
			"conditions":
			{
				"selector": ".gallery-main-images a img",
				"key": "src"
			}
		}
	}
}
Leave a Comment