serpapi / public-roadmap Goto Github PK
View Code? Open in Web Editor NEWPublic roadmap for SerpApi, LLC (https://serpapi.com)
Public roadmap for SerpApi, LLC (https://serpapi.com)
The Yandex search is returning the url
and thumbnail
for the inline_images
without base domain.
https://serpapi.com/playground?engine=yandex&text=spider+man&no_cache=true
X-Ray is highlighting element on the top and bottom carousels. It shouldn't highlight the bottom carousel element.
The position is also reset for the bottom carousel and X-Ray doesn't work properly
Inspect: https://serpapi.com/searches/61c112a796f5d71b1832e169/inspect
This block shows up when you add tbs:local_avail:1 parameter.
Screenshot:
Google Local Pack, and Google Local Results have problematic results where addresses are sometimes parsed as numbers etc.
Employing Machine Learning classification models to improve the parser would be a viable solution.
Related Blog Posts About the Process:
#1) How to scrape Google Local Results with Artificial Intelligence?
#2) Real World Example of Machine Learning on Rails
#3) AI Training Tips and Comparisons
#4) Machine Learning in Scraping with Rails
#5) Implementing ONNX models in Rails
#6) How ML Hybrid Parser Beats Traditional Parser
#7) How to Benchmark ML Implementations on Rails
The date
doesn't seem to be extracted anymore on our top_stories
block on Google regular.
Playground: https://serpapi.com/playground?q=Inflation
Inspect: https://serpapi.com/searches/6212dce866440aa54c47b955/inspect
Original report:
Ref: https://forum.serpapi.com/bugs/p/date-missing-from-topstories
Inspect: https://serpapi.com/searches/6081db17d5a5313c7cc3cfb9/inspect
On Google Shopping results, if the result is also listed on Google Products, it will include an element that shows how many listings would be on the product page:
Past Canny Conversation:
Angelo: Emirhan Akdeniz, great! What's the ETA on that?
Emirhan Akdeniz: We are trying to separate Google Shopping as another engine. It would enable us to solve the issues much faster and effectively in the future. So far, every issue we heard from customers about Google Shopping, including this issue has been fixed in a pull request. We are reviewing last state of the update. My estimates on ETA is 1-2 weeks since it is a big review process. Thanks so much for your enthusiasm and patience.
Angelo: thank you!!
An user has noticed that the Google url has changed from chips
to htchips
. And the filters are not working anymore otherwise:
Justin tried to add this to the playground but it did not work because we don't support that parameter:
Original Canny issue: https://forum.serpapi.com/bugs/p/google-jobs-api-update-chips-parameter-to-htchips
This is a different block that needs to be parsed under another key.
Playground Link
Inspect Link
Add support for population answers. e.g. "population of China"
Playground: https://serpapi.com/playground?engine=baidu&q=%E4%B8%AD%E5%9B%BD%E4%BA%BA%E5%8F%A3
Google has a What to watch
section in their search results and currently we aren't return data for it in our API.
Inspect: https://serpapi.com/searches/6225b8e61c450a597413aab4/inspect
More than a few users are looking to manually (not automatically) renew early their account.
If initiating this process is not foreseen to be creating such problems that we have had at November 5, 2021 billing issue then this will be very convenient for some of our users.
Please feel free to discuss but if no blocker then next for sure.
Google Shopping Results sometimes give "Searches Related To" part which could be helpful for customers.
Inspect Link
One of our users reached out to report that IM parameter of Baidu search is currently not working.
The Playground:
https://serpapi.com/playground?engine=baidu&q=Coffee&lm=7
The Working Version:
Guidance from our user:
I got that url by...
going to the url generated by your playground as you had set it up
clicking on the filter tools (look for the little filter icon on the right under the search bar)
the tools are dropdown menus and are revealed just as in google when you click on 'tools'
the left-most dropdown is the one that controls the time.
I set that to 'within 7 days', which is '一周内', the second option
that successfully limited the results to the time specified and the first results page had the url above when I did this
By experimenting I found that the arguments that vary as I change the time limit are 'rsv_t' and 'gpc', but obviously the value of these parameters is encoded somehow.
next as we have received a couple of requests about this.
Welp, it seems like the only thing keeping SerpAPI from turning into a Super Saipan is Yelp!
But seriously, it would be nice to be able to scrape Yelp search results. Whether I like it or not, Yelp has the best combination of business names, reviews, and photos, and plenty of people trust what users on Yelp say. It’s a must for SerpAPI!
Walmart Product API returns empty results. Checked twice, those products return results but more than occasional hiccups preventing user completing their task.
Product ID: 913713226
Inspect: https://serpapi.com/searches/622117892236e2bcc473de66/inspect
{
"search_metadata": {
"id": "622117892236e2bcc473de66",
"status": "Error",
"json_endpoint": "https://serpapi.com/searches/d802734da53bc2d7/622117892236e2bcc473de66.json",
"created_at": "2022-03-03 19:31:21 UTC",
"processed_at": "2022-03-03 19:31:21 UTC",
"walmart_product_url": "https://www.walmart.com/ip/913713226?selected=true",
"total_time_taken": 450.98
},
"search_parameters": {
"product_id": "913713226",
"engine": "walmart_product",
"device": "desktop"
},
"error": "We couldn't get valid results for this search. Please try again later. Set the no_cache parameter to true in order to force a retry."
}
The screenshot:
The Playground: https://serpapi.com/playground?engine=walmart_product&product_id=913713226
User also stated that they had problem with 994807896 but Playground seems fine:
https://serpapi.com/playground?engine=walmart_product&product_id=994807896
next.
The HTML in the Yandex Images engine doesn't work
https://serpapi.com/playground?engine=yandex_images&text=spider+man&no_cache=true
One of our users reported that Google does not parse several elements from certain mobile results page. The differences here are that search device is mobile ad we are searching for videos of a darling of TikTok users: Charli D'amelio.
Missing elements in the results: (All numbered in the screenshot)
1- Video tags and their links
2- Youtube channel's link and a selection of videos
3- Short videos
4- People also search for (contents of this element is returned in related_searches key)
5- TikTok channel's link and a selection of videos
6- TikTik category link and a selection of videos
7- Broaden this search
8- Related searches (content of this element is not returned at all, returned content is from people also search for section)
The Playground:
The Inspect:
https://serpapi.com/searches/6213ce7f49ecdb4ebc4202ed/inspect
Screenshot:
The first product in SerpApi JSON got the wrong source into source
field in Google Shopping API. CSS seems be leaking.
Inspect: https://serpapi.com/searches/6216b3680ead41878b9fdd9d/inspect
Google Ad Results API is missing phone number extraction from the "top"
"block_position"
ad results.
For more, have a look at the playground link or json_endpoint
link.
Add support for direct answers. e.g. "How high is mount Fuji?"
Playground: https://serpapi.com/playground?engine=baidu&q=%E5%AF%8C%E5%A3%AB%E5%B1%B1%E6%9C%89%E5%A4%9A%E9%AB%98
Google Shopping API no results for "dog supplies" and limited results for other basic queries
Playground:
https://serpapi.com/playground?q=dog+supplies&location=Austin%2C+Texas%2C+United+States&gl=us&hl=en&tbm=shop
Inspect:
https://serpapi.com/searches/61d7332bc47d3ca7b1f3d6b2/inspect
The Playground: (This returns results to me)
Inspect examples:
https://serpapi.com/searches/61d82dad5f220316258218a3/inspect
https://serpapi.com/searches/61d82da5aa6c95fba3174ddd/inspect
https://serpapi.com/searches/61d82edf8cd8782812cb365f/inspect
Screenshot:
Inspect: https://serpapi.com/searches/61d82facaef2646991fb90dc/inspect
Screenshot:
Another example:
https://serpapi.com/searches/61dd5070b7b1cc879f1b71ff/inspect
Another example:
https://serpapi.com/searches/6202cdc4d5a53137721bb4b8/inspect
Another example: https://serpapi.com/searches/62091327c47d3cbb63618531/inspect
Google Search API inline_video_carousels
is missing first channel link
Playground: https://serpapi.com/playground?q=addison+rae&location=Austin%2C+Texas%2C+United+States&gl=us&hl=en
Inspect: https://serpapi.com/searches/6213c8b749ecdb4f6252bde7/inspect
Another one (but currently seems to be working):
SerpApi parser replaces the first link in the container with this Youtube link.
Ref: Google Search API inline_video_carousels Missing First Channel Link
Add support for calculator answers. e.g. "2+2"
Playground: https://serpapi.com/playground?engine=baidu&q=2%2B2
This repository is taking over from our old forum (https://forum.serpapi.com/):
We'll be migrating past reports from https://forum.serpapi.com/ to this repository. Feel free to re-add your issue if we haven't already.
Add support for weather answers. e.g. "What's the weather like in Beijing?"
It would be useful to include the simg code necessary to do an "All Sizes" similar image search in the json results for a reverse image search. Right now, the html version of the reverse image search includes the link to the appropriate search, but it is not included in the JSON, which makes automating that extra step difficult.
`
One of our users reached out to report that Yahoo Shopping API does not work properly when sent certain characters.
The Playground Query: Great Value 2% Reduced Fat Milk, 128 Fl Oz - No Result
The Playground Query: Great Value 2 Reduced Fat Milk, 128 Fl Oz - No Result
https://serpapi.com/playground?engine=yahoo_shopping&p=Great+Value+2+Reduced+Fat+Milk%2C+128+Fl+Oz
The Playground Query: Great Value 2 Reduced Fat Milk 128 Fl Oz - Successful when removed comma and percentage
https://serpapi.com/playground?engine=yahoo_shopping&p=Great+Value+2+Reduced+Fat+Milk+128+Fl+Oz
The user submitted video:
Inspect: https://serpapi.com/searches/6225c5858b7ed2c74b43ccdd/inspect
next.
Google Maps Review API does not always sort by highest rated reviews first:
data_id=0x3b080dafbed183bf:0x5951f316ba13a37e
The Playground: https://serpapi.com/playground?engine=google_maps_reviews&data_id=0x3b080dafbed183bf%3A0x5951f316ba13a37e&hl=en&sort_by=ratingHigh&no_cache=true
The Inspect: https://serpapi.com/searches/6229bd6f37b24ae465097169/inspect
I have described it as not working properly because it outputs blank results when it does not and it tends to work for some places and not for others:
The default on the Playground works for example:
data_id=0x89c259a61c75684f:0x79d31adb123348d2
Inspect: https://serpapi.com/searches/6229c9c67690dcd0a6b5c112/inspect
Another random example data_id=0x89c258ff0f7a7793:0xbedc0a7158ebea16
This is working.
next.
Separate and make another engine out of tbm=shop
parameter for the ease of use, development, and reviewing.
In Google Shopping Results tag
is repeated within extensions
for some results.
When using tbs
parameter in the playground its value is lowercased. tbs
value is case sensitive and should not be lowercased, otherwise, the parameter will have no effect.
Add support for finance answers. e.g. "Google stock"
Playground: https://serpapi.com/playground?engine=baidu&q=%E8%B0%B7%E6%AD%8C%E8%82%A1%E7%A5%A8
The Baidu engine organic result has a rich_snippet
now it returns empty extensions
for some results:
https://serpapi.com/playground?engine=baidu&q=elon+musk&ct=2&no_cache=true
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.