This PR adds an implementation of Azure Computer Vision that allows the bot to attempt descriptions of images in English. An image description is triggered by the keyword !describe (in addition to the comment mentioning the bot's Reddit account).
Examples of some attempted descriptions can be seen in this Reddit thread. The bot describes some images very well, and others at an "okay" level. Some are not that great, and in some cases, the bot does not even attempt to give a transcription.
Note that currently the implementation is not very robust. Pointing to a non-existent image is likely to cause the bot to crash (I haven't tried 😑). Depending on availability, I may be able to address this before tonight's release. (Or maybe not)
This PR adds an implementation of Azure Computer Vision that allows the bot to attempt descriptions of images in English. An image description is triggered by the keyword
!describe(in addition to the comment mentioning the bot's Reddit account).Examples of some attempted descriptions can be seen in this Reddit thread. The bot describes some images very well, and others at an "okay" level. Some are not that great, and in some cases, the bot does not even attempt to give a transcription.
Note that currently the implementation is not very robust. Pointing to a non-existent image is likely to cause the bot to crash (I haven't tried 😑). Depending on availability, I may be able to address this before tonight's release. (Or maybe not)
Closes #83