Spaces:
Paused
Paused
try combining into one app
Browse files
app.py
CHANGED
|
@@ -14,7 +14,7 @@ from langchain_community.document_loaders import DataFrameLoader
|
|
| 14 |
from langchain.text_splitter import RecursiveCharacterTextSplitter
|
| 15 |
from langchain_community.vectorstores import FAISS
|
| 16 |
from langchain.chains import RetrievalQA
|
| 17 |
-
from tokopedia import request_product_id, request_product_review
|
| 18 |
import gradio as gr
|
| 19 |
|
| 20 |
shop_id = ""
|
|
@@ -34,6 +34,138 @@ logging.basicConfig(
|
|
| 34 |
logger = logging.getLogger(__name__)
|
| 35 |
|
| 36 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 37 |
def scrape(product_id, max_reviews=LIMIT):
|
| 38 |
all_reviews = []
|
| 39 |
page = 1
|
|
|
|
| 14 |
from langchain.text_splitter import RecursiveCharacterTextSplitter
|
| 15 |
from langchain_community.vectorstores import FAISS
|
| 16 |
from langchain.chains import RetrievalQA
|
| 17 |
+
# from tokopedia import request_product_id, request_product_review
|
| 18 |
import gradio as gr
|
| 19 |
|
| 20 |
shop_id = ""
|
|
|
|
| 34 |
logger = logging.getLogger(__name__)
|
| 35 |
|
| 36 |
|
| 37 |
+
import requests
|
| 38 |
+
|
| 39 |
+
|
| 40 |
+
def request_product_id(shop_domain, product_key):
|
| 41 |
+
endpoint = "https://gql.tokopedia.com/graphql/PDPGetLayoutQuery"
|
| 42 |
+
payload = {
|
| 43 |
+
"operationName": "PDPGetLayoutQuery",
|
| 44 |
+
"variables": {
|
| 45 |
+
"shopDomain": f"{shop_domain}",
|
| 46 |
+
"productKey": f"{product_key}",
|
| 47 |
+
"apiVersion": 1,
|
| 48 |
+
},
|
| 49 |
+
"query": """fragment ProductVariant on pdpDataProductVariant {
|
| 50 |
+
errorCode
|
| 51 |
+
parentID
|
| 52 |
+
defaultChild
|
| 53 |
+
children {
|
| 54 |
+
productID
|
| 55 |
+
}
|
| 56 |
+
__typename
|
| 57 |
+
}
|
| 58 |
+
|
| 59 |
+
query PDPGetLayoutQuery($shopDomain: String, $productKey: String, $layoutID: String, $apiVersion: Float, $userLocation: pdpUserLocation, $extParam: String, $tokonow: pdpTokoNow, $deviceID: String) {
|
| 60 |
+
pdpGetLayout(shopDomain: $shopDomain, productKey: $productKey, layoutID: $layoutID, apiVersion: $apiVersion, userLocation: $userLocation, extParam: $extParam, tokonow: $tokonow, deviceID: $deviceID) {
|
| 61 |
+
requestID
|
| 62 |
+
name
|
| 63 |
+
pdpSession
|
| 64 |
+
basicInfo {
|
| 65 |
+
id: productID
|
| 66 |
+
}
|
| 67 |
+
components {
|
| 68 |
+
name
|
| 69 |
+
type
|
| 70 |
+
position
|
| 71 |
+
data {
|
| 72 |
+
...ProductVariant
|
| 73 |
+
__typename
|
| 74 |
+
}
|
| 75 |
+
__typename
|
| 76 |
+
}
|
| 77 |
+
__typename
|
| 78 |
+
}
|
| 79 |
+
}
|
| 80 |
+
""",
|
| 81 |
+
}
|
| 82 |
+
|
| 83 |
+
headers = {
|
| 84 |
+
"User-Agent": "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/118.0.0.0 Safari/537.36",
|
| 85 |
+
"Referer": "https://www.tokopedia.com",
|
| 86 |
+
"X-TKPD-AKAMAI": "pdpGetLayout",
|
| 87 |
+
}
|
| 88 |
+
|
| 89 |
+
return requests.request(
|
| 90 |
+
method="POST", url=endpoint, json=payload, headers=headers, timeout=30
|
| 91 |
+
)
|
| 92 |
+
|
| 93 |
+
|
| 94 |
+
def request_product_review(product_id, page=1, limit=20):
|
| 95 |
+
ENDPOINT = "https://gql.tokopedia.com/graphql/productReviewList"
|
| 96 |
+
payload = {
|
| 97 |
+
"operationName": "productReviewList",
|
| 98 |
+
"variables": {
|
| 99 |
+
"productID": f"{product_id}",
|
| 100 |
+
"page": page,
|
| 101 |
+
"limit": limit,
|
| 102 |
+
"sortBy": "",
|
| 103 |
+
"filterBy": "",
|
| 104 |
+
},
|
| 105 |
+
"query": """query productReviewList($productID: String!, $page: Int!, $limit: Int!, $sortBy: String, $filterBy: String) {
|
| 106 |
+
productrevGetProductReviewList(productID: $productID, page: $page, limit: $limit, sortBy: $sortBy, filterBy: $filterBy) {
|
| 107 |
+
productID
|
| 108 |
+
list {
|
| 109 |
+
id: feedbackID
|
| 110 |
+
variantName
|
| 111 |
+
message
|
| 112 |
+
productRating
|
| 113 |
+
reviewCreateTime
|
| 114 |
+
reviewCreateTimestamp
|
| 115 |
+
isReportable
|
| 116 |
+
isAnonymous
|
| 117 |
+
reviewResponse {
|
| 118 |
+
message
|
| 119 |
+
createTime
|
| 120 |
+
__typename
|
| 121 |
+
}
|
| 122 |
+
user {
|
| 123 |
+
userID
|
| 124 |
+
fullName
|
| 125 |
+
image
|
| 126 |
+
url
|
| 127 |
+
__typename
|
| 128 |
+
}
|
| 129 |
+
likeDislike {
|
| 130 |
+
totalLike
|
| 131 |
+
likeStatus
|
| 132 |
+
__typename
|
| 133 |
+
}
|
| 134 |
+
stats {
|
| 135 |
+
key
|
| 136 |
+
formatted
|
| 137 |
+
count
|
| 138 |
+
__typename
|
| 139 |
+
}
|
| 140 |
+
badRatingReasonFmt
|
| 141 |
+
__typename
|
| 142 |
+
}
|
| 143 |
+
shop {
|
| 144 |
+
shopID
|
| 145 |
+
name
|
| 146 |
+
url
|
| 147 |
+
image
|
| 148 |
+
__typename
|
| 149 |
+
}
|
| 150 |
+
hasNext
|
| 151 |
+
totalReviews
|
| 152 |
+
__typename
|
| 153 |
+
}
|
| 154 |
+
}
|
| 155 |
+
""",
|
| 156 |
+
}
|
| 157 |
+
|
| 158 |
+
headers = {
|
| 159 |
+
"User-Agent": "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/118.0.0.0 Safari/537.36",
|
| 160 |
+
"Referer": "https://www.tokopedia.com",
|
| 161 |
+
"X-TKPD-AKAMAI": "productReviewList",
|
| 162 |
+
}
|
| 163 |
+
|
| 164 |
+
return requests.request(
|
| 165 |
+
method="POST", url=ENDPOINT, json=payload, headers=headers, timeout=30
|
| 166 |
+
)
|
| 167 |
+
|
| 168 |
+
|
| 169 |
def scrape(product_id, max_reviews=LIMIT):
|
| 170 |
all_reviews = []
|
| 171 |
page = 1
|