Wondering if we could start another thread regarding your javascript / Chrome serial api driver / library? Will you be making that available on Github or other means? Could be very useful to other developers.
Yes of course, it was my intention. I'll make my code roughly presentable and do that asap. (At this time it's more a feasibility study, and I did found not a way to stream data to file. Even with the chrome filesystem api, user seems to have to make an action. The only solution I see is to make a simple node.js server, to store to file or database.)
Ah, I see now rereading your post that the Chrome fileSystem API is not cooperating with you, likely having to do with the sandboxing. Did find a couple related links,
Thanks for the links, but as said, this is what I already explored. I can save to file or save to datastore : in the current version data is buffered in memory and can then be saved to a file, but the user has to click to save the file. In all cases, even with the 'no user action' solutions to datastore, it is a sanboxed access to quota-limited store (typically 50Mo max) that will not fit the needs.
I started a node.js version using node-serialport and node-ws.
> In all cases, even with the 'no user action' solutions to datastore, it is a sandboxed access to quota-limited store (typically 50Mo max) that will not fit the needs.
Am I reading the http://danml.com/download.html link incorrectly? I had the impression that this is a way out of the sandbox, and possibly the quota as well. Still though looks like it requires some user clicks.
Your node.js version does sound like it avoids all these limitations.
If the javascript program is only doing neurofeedback or neurogaming (without the need to save the entire raw session data) -- then various signal processing libraries may open possibilities,
Yes I pushed a basis for the node.js version some times ago, I am currently on merging the webapp and the chromeapp to avoid duplicate « client side » code. I wanted to change from current time-based protocol to a parser based one that would allow to decode board state, but I am facing a problem : I got a lot of errors that do not allow the parser to work well, and I currently don't know if these are coming from board, from drivers or from elsewhere.
For example, I sometimes got in my logs (directly from node-serial output) :
@willemmulder ;: I use linux, so I didn't notice this issue. Does it work on Windows with the chrome app serial ? If so, I could modify the app to stream data from chrome app to node server through websocket instead of using serialportnode.
@biomurph In my configuration, the errors are quite reproductible, especially for the iWedance, I plan to compare logs with the chrome app version as soon as the settings management will be implemented.
(but I save these changes and tests for later, I'm currently making some test with the cloud signal processing and cognitive computing api, and as it's a time limited trial, I shifted my priorities )
Comments
I started a node.js version using node-serialport and node-ws.
is a sandboxed access to quota-limited store (typically 50Mo max) that
will not fit the needs.
Am I reading the http://danml.com/download.html link incorrectly? I had the impression that this is a way out of the sandbox, and possibly the quota as well. Still though looks like it requires some user clicks.
Your node.js version does sound like it avoids all these limitations.
https://www.google.com/search?q=javascript+signal+processing
https://www.google.com/search?q=javascript+ui+framework+comparison+2014
For example, I sometimes got in my logs (directly from node-serial output) :
« updating iWedance detect setings »
instead of
« updating impedance detect setings »
or
« CH6SET, 0x0A, 0x69, 0, 1, 1, 0, 1, 0, 0, 1
, 0x0B, 0x69, 0, 1, 1, 0, 1, 0, 0, 1
CH8SET, 0x0C, 0x69, 0, 1, 1, 0, 1, 0, 0, 1 »
instead of
« CH6SET, 0x0A, 0xE0, 1, 1, 1, 0, 0, 0, 0, 0
CH7SET, 0x0B, 0x68, 0, 1, 1, 0, 1, 0, 0, 0
CH8SET, 0x0C, 0x68, 0, 1, 1, 0, 1, 0, 0, 0 »
and so on, so i am investigating further for now.
@biomurph
In my configuration, the errors are quite reproductible, especially for the iWedance, I plan to compare logs with the chrome app version as soon as the settings management will be implemented.
(but I save these changes and tests for later, I'm currently making some test with the cloud signal processing and cognitive computing api, and as it's a time limited trial, I shifted my priorities