Collect postcards locally?

[fishy]

As I understand it, the current blueprint to auto collect postcards collect them on birdbuddy. Even though the videos and images are viewable from HA's media integration, that's by implementing a media source from bird buddy integration, not really locally saved media files.

It would be great if we can collect the videos and images locally, to a mounted nas directory, for example. The media source integration seems to have some limit on how many videos/images you can see and older ones are no longer available there.

[jhansche]

Yeah, there are some sample automations in the forum that others have shared, that will download media to local files. Basically I'm an automation for the postcard event, you loop through the media and use the download integration to save them locally. I'll see if I can find the forum posts for it

[jhansche]

Here it is (and other ideas there also): https://community.home-assistant.io/t/custom-component-bird-buddy-smart-bird-feeder/497464/17?u=madcoder

[jhansche]

And a similar example showing how to loop over the media items: https://community.home-assistant.io/t/custom-component-bird-buddy-smart-bird-feeder/497464/66?u=madcoder

[jhansche]

Sorry, I'm thinking in terms of the python library, as that's what those links were for. But this is the HA repo 😳

In HA (specifically once it gets translated into the event system), it's just a plain data dictionary, not the pybirdbuddy wrappers.

Instead of sighting.species.name, you would use dictionary syntax: sighting["species"]["name"]

[fishy]

weird, even with sighting["species"]["name"]' I'm still getting 'dict object' has no attribute 'species'`:

[jhansche]

In the traces UI you're showing here, select the trigger node (*) at the top, and then click the "Changes variables" tab. It'll list the full yaml (a superset of json) of the trigger data. From here you should be able to spot the species data

Keep in mind that it needs real data, so if you're testing it by clicking "Run actions" in the automation editor, it's not going to work.

From here, it's actually not the trigger-level sighting that has the data, but sightingReport (you were close, above, but the name is not report):

{{ trigger.event.data.sighting.sightingReport.sightings[0].species.name }}

But this might not be present, if BB did not identify the species with confidence. In the case of unidentified, you'll just get the list of potential matches. This is what the integration handles for you, with strategy=mystery or best-guess. But it handles that when collecting the postcard, not in the event data.

[fishy]

After some trial and error, in the end this is the setup I'm happy with:

sequence:
  - alias: Collect a Postcard
    data_template:
      strategy: mystery
      best_guess_confidence: 20
      share_media: false
      postcard: "{{ trigger.event.data.postcard }}"
      sighting: "{{ trigger.event.data.sighting }}"
    enabled: true
    action: birdbuddy.collect_postcard
  - alias: Download images
    repeat:
      count: "{{ trigger.event.data.sighting.medias | length }}"
      sequence:
        - action: downloader.download_file
          metadata: {}
          data:
            overwrite: true
            url: >-
              {{ trigger.event.data.sighting.medias[repeat.index - 1].contentUrl
              }}
            filename: >-
              {{ as_timestamp(trigger.event.data.postcard.createdAt) |
              timestamp_custom('%Y-%m-%d_%H-%M-%S')
              }}-{{trigger.event.data.sighting.sightingReport.sightings[0].species.name.replace("
              ", "_")}}-{{ "%02d" % repeat.index }}.jpg
            subdir: >-
              birdbuddy/{{ as_timestamp(trigger.event.data.postcard.createdAt) |
              timestamp_custom('%Y-%m-%d') }}
  - alias: Download video
    action: downloader.download_file
    metadata: {}
    data:
      overwrite: true
      url: "{{ trigger.event.data.sighting.videoMedia.contentUrl }}"
      filename: >-
        {{ as_timestamp(trigger.event.data.postcard.createdAt) |
        timestamp_custom('%Y-%m-%d_%H-%M-%S')
        }}-{{trigger.event.data.sighting.sightingReport.sightings[0].species.name.replace("
        ", "_")}}.mp4
      subdir: >-
        birdbuddy/{{ as_timestamp(trigger.event.data.postcard.createdAt) |
        timestamp_custom('%Y-%m-%d') }}

Some notes:

1. It can save both video and images
2. I used the postcard createdAt instead of the individual medias' createdAt as the source of the timestamp in the filename, this way all the images/video from the same postcard will have the same timestamp in the filename and be less messy
3. I also padded the index of the images with 0, as sometimes a postcard can have >10 images so this way they sort better
4. I also group them in subdirectories by date
5. Species name is in the filename, with space replaced by underscore

[jhansche]

Nice work 🙂