I guess that’s relevant lol, sorry. The script is sending the results to Devonthink as clutter free PDF files. So there can easily be a “backlog” of conversions if there are several search sets lined up. I can watch the DT database populate with the results. As it does, it adds a new group for each search set. A new group is never added until the previous group is complete.
So if there are 6 searches executed at 12:00.
As soon as one of the searches completes, the DT group is created and the results added as PDFs
So, there is a period of time where all six searches are complete, but DT is still processing the results of search #3. I am wondering if technically, that means the search set itself is still considered active by DA and that’s why it sometimes “skips” the 12:15 run when DT gets too far behind.
I guess my question is this – when the searches are done, but DT has not yet processed the results per the script, where are those pending results? Are they still in DA or in a system cache or in DT cache? Maybe I simply need a line or two added to my script to tell it to “let DA go” or something??
I apologize for this whole thread, i know it’s been all over the place.
No worries!
@cgrunenberg could answer more authoritatively, but I believe it would still be in DEVONagent’s hands.
The script doesn’t matter as it’s executed asynchronously in the background but as the order of multiple scheduled searches performed concurrently is undefined (e.g. depending on the time of the last run and the duration of each search), it might be better to use different times and/or to merge the search sets.
In addition, scheduled searches might reuse existing and unused search windows, therefore 6 scheduled searches do not necessarily open 6 windows. E.g. if one scheduled & completed search does not return any results (e.g. due to settings or cache) and nothing gets logged, then another search will reuse this window.
1 Like
Thanks @cgrunenberg. I am happy to report that after setting the QUIT option on in the search sets and seriously thinning the search volumes (including no link follows) it seems to be working consistently now.
One final question (I hope). My only reason for not wanting to combine search sets is that the script assigns them to groups in DT based on the search. If I was doing car maintenance research, I would like to have separate groups in DT for TIRES, BATTERY, ELECTRICAL SYSTEM, ENGINE etc. If I put them all together in one big “CAR” group, it would really screw up my workflow within DT, that exports selected documents to TheBrain app. Now, I can change it all up and try to have DT’s AI do the classifying for me, but it’s not ideal due to the heavy overlap in content between searches.
So if I play around with scripting do you think there is a way to assign a “tag” to the files to show which keyword triggered it so i can use search set of “tires OR battery OR electrical system”?
So if I play around with scripting do you think there is a way to assign a “tag” to the files to show which keyword triggered it so i can use search set of “tires OR battery OR electrical system”?
It’s quite possible the text of sites from individual searches, e.g., one for tires, one for battery, etc., could contain text from the other searches, so it may not be easily done by checking the text of results.
