v2 mongodb version

.gitignore

@@ -1,2 +1 @@
-PHPSESSID.txt
+.env
-download

README.md

@@ -1,11 +0,0 @@
-This is a python script for downloading original pixiv images from popular search results via a premium account.
-
-# Instructions
-
-1. Download this repo to your computer of course, and open the terminal in it. Run `pip install -r requirements.txt` if necessary.
-
-2. In your browser, on Pixiv logged in to a premium account, in dev tools Application tab, copy the **value** of the `PHPSESSID` cookie, and paste it into a new file named `PHPSESSID.txt` in this folder.
-
-3. Run `python pixiv-popular-downloader.py -h` for usage information. Example usage to download 10 pages of 初音ミク tag, including r18: `python pixiv-popular-downloader.py -r -p 10 "初音ミク"`
-
-4. Check the download folder. If you're getting newest results instead of popular results, then your PHPSESSID failed to work.

pixiv-popular-downloader.py

@@ -3,49 +3,145 @@ import requests
 from requests_toolbelt.adapters import host_header_ssl
 from urllib.parse import quote as encodeURI
 import os
+from pymongo import MongoClient
+from gridfs import GridFS
+from datetime import datetime
+from dotenv import load_dotenv
+load_dotenv()
 
 ap = argparse.ArgumentParser()
 ap.add_argument("tag", help="Pixiv tag(s) to search")
 ap.add_argument("-p", dest="numpages", type=int, default=1, help="number of pages to download (default 1)")
 ap.add_argument("-s", dest="startpagenum", type=int, default=1, help="page number to start at")
-ap.add_argument("-r", action='store_true', help="include r18 posts")
 args = ap.parse_args()
 
-PHPSESSID = None
-with open("PHPSESSID.txt", 'r') as f:
-	PHPSESSID = f.read()
 
 rqs = requests.Session()
 rqs.mount('https://', host_header_ssl.HostHeaderSSLAdapter())
 
-download_count = 1
+
-for i in range(args.startpagenum, args.numpages+1):
+
-	page_url = f"https://210.140.131.219/ajax/search/artworks/{encodeURI(args.tag, safe='')}?order=popular_d&mode={'all' if args.r else 'safe'}&p={i}"
+dbclient = MongoClient(os.environ["MONGODB_URI"])
-	print("get", page_url)
+
-	page_data = rqs.get(page_url, cookies={"PHPSESSID": PHPSESSID}, headers={"host":"www.pixiv.net"}).json()
+db = dbclient["mikudb"]
-	if (page_data['error']):
+illustration_collection = db["illustration_collection"]
-		print(page_data['message'])
+search_collection = db["search_collection"]
-		exit(1)
+
-	for illust in page_data['body']['illustManga']['data']:
+gridfs = GridFS(db)
-		illust_r18 = bool(illust['xRestrict'])
+
-		illust_url = f"https://210.140.131.219/ajax/illust/{illust['id']}/pages"
+
-		print("get", illust_url)
+
-		illust_data = rqs.get(illust_url, headers={"host":"www.pixiv.net"}).json()
+
-		if (illust_data['error']):
+def download_popular(tag, startpagenum = 1, numpages = 1):
-			print(illust_data['message'])
+	
-		else:
+
-			for image in illust_data['body']:
+	# record this search session in the database
-				image_url = image['urls']['original']
+	search_document_id = search_collection.insert_one({
-				download_dir = f"download/{args.tag}/"
+		"date": datetime.now(), # date started
-				os.makedirs(download_dir, exist_ok=True)
+		"query": None, # the tag being searched
-				download_filename = str(download_count) + '_' + ('x_' if illust_r18 else '') + image_url.split('/').pop()
+		"current_page": startpagenum, # keep track of the page we're on
-				download_path = download_dir + download_filename
+		"current_illust": None, # keep track of which item is being downloaded
-				if os.path.exists(download_path):
+		"search_data": [], # save each payload
-					print(download_path, "already exists")
+		"results": [], # collect ids of all results that were saved
-					continue
+		"completed": False # whether this ever reached the end
-				print("get", image_url)
+	}).inserted_id
-				res = rqs.get(image_url, headers={'referer':'https://www.pixiv.net'})
+
-				with open(download_path, "wb") as f:
+
-					f.write(res.content)
+	the_id_of_the_first_result_on_the_previous_page = None
-				print("saved", download_filename)
+	for page_number in range(startpagenum, numpages+1):
-				download_count = download_count + 1
+		# download search results
+		search_url = f"https://210.140.131.219/ajax/search/artworks/{encodeURI(tag, safe='')}?order=popular_d&mode=all&p={page_number}"
+		print("get", search_url)
+		search_data = rqs.get(search_url, cookies={"PHPSESSID": os.environ["PHPSESSID"]}, headers={"host":"www.pixiv.net"}).json()
+		
+		# save raw search data
+		search_collection.update_one({"_id": search_document_id}, {"$set": {"current_page": page_number}, "$push": {"search_data": {
+			"page_number": page_number,
+			"search_url": search_url,
+			"search_data": search_data
+		}}})
+
+		if (search_data['error']):
+			print("error from search api:", search_data['message'])
+			exit(1)
+
+		search_results = search_data['body']['illustManga']['data']
+
+		# if there is no data then we are done
+		if not search_results:
+			print("No more search results")
+			search_collection.update_one({"_id": search_document_id}, {"$set":{"completed": True}})
+			exit()
+
+		# But large tags seem to give the last page of results for any page number
+		if search_results[0]['id'] == the_id_of_the_first_result_on_the_previous_page:
+			print("Reached duplicate search results, looks like the end")
+			search_collection.update_one({"_id": search_document_id}, {"$set":{"completed": True}})
+			exit()
+		else: the_id_of_the_first_result_on_the_previous_page = search_results[0]['id']
+
+
+		# for each search result
+		for illust_data_from_search in search_results:
+			illust_id = illust_data_from_search['id']
+
+			search_collection.update_one({"_id": search_document_id}, {"$set":{"current_illust": illust_id}})
+
+			# check if this illust was already saved
+			if illustration_collection.count_documents({"_id": illust_id}):
+				print("already have", illust_id)
+				continue
+
+			# illust_metadata_from_search has limited information (blank descriptions)
+			# download full meta data from ajax
+			illust_ajax_url = f"https://210.140.131.219/ajax/illust/{illust_id}"
+			print("get", illust_ajax_url)
+			illust_ajax_data = rqs.get(illust_ajax_url, cookies={"PHPSESSID": os.environ["PHPSESSID"]}, headers={"host":"www.pixiv.net"}).json()
+			if (illust_ajax_data['error']):
+				print("error from ajax api:", illust_ajax_data['message'])
+
+			# illust_ajax_data does not have "page" data (additional image urls)
+			# download that
+			illust_pages_url = illust_ajax_url + "/pages"
+			print("get", illust_pages_url)
+			illust_pages_data = rqs.get(illust_pages_url, cookies={"PHPSESSID": os.environ["PHPSESSID"]}, headers={"host":"www.pixiv.net"}).json()
+			if (illust_pages_data['error']):
+				print("error from ajax pages api:", illust_pages_data['message'])
+
+			else:
+				
+				# prepare database document
+				document = {
+					"_id": illust_id, # use the unique artwork id for document id so we can't have duplicates
+					"illust_ajax_data": illust_ajax_data, # save all the metadata for the artwork
+					"illist_pages_data": illust_pages_data, # save all the image urls of the data
+					"downloaded_images": {}, # map of image filenames to gridfs ids
+					"date_saved": datetime.now()
+				}
+				
+				# download originals
+				for illust_page_data in illust_pages_data['body']:
+
+					original_image_url = illust_page_data['urls']['original']
+					original_image_filename = original_image_url.split('/').pop()
+
+					print("get", original_image_url)
+					res = rqs.get(original_image_url, headers={'referer':'https://www.pixiv.net'})
+
+					print("gridfs put", res.url)
+					gridfs_id = gridfs.put(res.content)
+					document['downloaded_images'][original_image_filename] = gridfs_id
+
+
+				# add to db
+				illustration_collection.insert_one(document)
+				search_collection.update_one({"_id": search_document_id}, {"$push": {"results":{
+					"id": illust_id
+				}}})
+
+	search_collection.update_one({"_id": search_document_id}, {"$set":{"completed": True}})
+	print("end of loop")
+
+
+download_popular(args.tag, args.startpagenum, args.numpages)
+

requirements.txt

@@ -1,3 +1,5 @@
 requests==2.26.0
 requests-toolbelt==0.9.1
 urllib3==1.26.6
+pymongo==3.12.0
+python-dotenv==0.19.0