Azure-Samples / Azure-Samples/cognitive-services-dotnet-sdk-samples
[Image Search API] It could not return plenty of results from multiple requests.
- Dominant language
- C#
- Stars
- 123
- Forks
- 166
- PR merge metrics
- No merged PRs in 30d
Description
## API Name (Kind)
`Image Search API`: `https://api.cognitive.microsoft.com/bing/v7.0/images/search`
## Issue Description
When I call Image Search API to collect images with a search term, say "car", I get only about 1000 results in total from all requests. I know there is a limitation for at most 150 results in each response, so I call the API multiple times with incremental offset. I don't understand which params I missed and I failed to find it from the official filter/option params list. Any idea for help? Thanks so much.
## Detail
The console log as below is from the script I wrote, which tries to collect 10000 image records with the search term "car". The offset comes from the `nextOffset` value in the API response. The request querying images from offset 1000 returns 87 records and the `nextOffset` is still 1000, which means the future request will be the same as the current. Thus my script comes to the exit. I tried several times even with other search terms but I never get a little more image records returned. Apparently, there must be much more images than a thousand for a general search term.
```
Searched 150 records from offset 0
Saved 121 new photos, now 121/10000
Searched 150 records from offset 173
Saved 142 new photos, now 263/10000
Searched 150 records from offset 345
Saved 149 new photos, now 412/10000
Searched 137 records from offset 537
Saved 137 new photos, now 549/10000
Searched 150 records from offset 719
Saved 138 new photos, now 687/10000
Searched 94 records from offset 892
Saved 93 new photos, now 780/10000
Searched 87 records from offset 1000
Saved 0 new photos, now 780/10000
Unable to collect more images. Done.
```
The query params in the request URL is like,
```
count=1000&freshness=month&imageType=photo&license=all&offset=0&q=car
```
Contributor guide
Assessment
This issue has not been assessed yet.