| [20:53:33] | <paul121[m]> | ACTION uploaded an image: (53KiB) < https://libera.ems.host/_matrix/media/r0/download/matrix.org/EXbxWaujDfO... > |
| [20:54:59] | <paul121[m]> | xc skiing amongst and under center pivots today 😁 |
| [20:56:50] | <paul121[m]> | Have a seasonal ski trail throughout your field.. special geometry edge case? |
| [21:23:58] | <symbioquine[m]> | <symbioquine[m]> "Alright, no worries. I'll see..." <- Interestingly, geoPHP is failing to detect the format of the WKB data and returning `FALSE`... (Hitting [this case](https://github.com/phayes/geoPHP/blob/685562416ec6d22b9b3927e02ca0ddacf8...)) |
| [21:24:04] | <symbioquine[m]> | Haven't figured out why yet. |
| [05:10:34] | <mstenta[m]> | > Oh yes they should! |
| [05:10:34] | <mstenta[m]> | symbioquine fyi i fixed these docs |
| [05:10:46] | <mstenta[m]> | <paul121[m]> "xc skiing amongst and under..." <- wow! how deep is that snow?? |
| [05:14:41] | <mstenta[m]> | <symbioquine[m]> "Interestingly, geoPHP is failing..." <- huh. weird. what is the first character of your WKB? based on that function it looks like it's expecting a `1`... |
| [05:14:45] | <mstenta[m]> | ACTION sent a code block: https://libera.ems.host/_matrix/media/r0/download/libera.chat/9ceda65da6... |
| [09:46:17] | * farmBOT has joined #farmos |
| [11:18:18] | <symbioquine[m]> | <mstenta[m]> "huh. weird. what is the first..." <- Interestingly, the first 11 bytes of the WKB I'm looking at right now are; `00 00 00 00 03 00 00 00 01 00 00` - which, based on the [WKT(B) wiki page](https://en.wikipedia.org/wiki/Well-known_text_representation_of_geometry...), means it's in big endian format... |
| [11:23:43] | <mstenta[m]> | Oh huh... I just ran a query on my local v1 db and they all seems to start with `01` |
| [11:23:46] | <mstenta[m]> | ACTION uploaded an image: (15KiB) < https://libera.ems.host/_matrix/media/r0/download/matrix.org/IvEEhdSomXN... > |
| [11:25:24] | <symbioquine[m]> | Care to share that query? 😄 |
| [11:25:58] | <mstenta[m]> | `SELECT * FROM field_data_field_farm_geofield;` |
| [11:26:23] | <mstenta[m]> | or `SELECT field_farm_geofield_geom FROM field_data_field_farm_geofield;` if you just want the geom column |
| [11:27:08] | <mstenta[m]> | (that screenshot was taken from PHPStorm's query console UI) |
| [11:27:54] | <symbioquine[m]> | Ah, must be PHPStorm doing the nice printing :( |
| [11:28:21] | <symbioquine[m]> | `SELECT HEX(field_farm_geofield_geom) FROM field_data_field_farm_geofield;` seems to work... |
| [11:28:23] | <mstenta[m]> | yep |
| [11:28:40] | <mstenta[m]> | ah there ya go |
| [11:29:53] | <mstenta[m]> | wonder if GeoPHP just doesn't support big endian? i did this search... maybe some clues? https://github.com/phayes/geoPHP/search?q=endian&type=issues |
| [11:38:08] | <symbioquine[m]> | Yeah, it looks like I only have 11 geometries in my DB with little endian WKB :( |
| [11:38:30] | <mstenta[m]> | Aw weird |
| [11:38:59] | <mstenta[m]> | And the big endian geometries show properly in v1 UI? |
| [11:39:00] | <symbioquine[m]> | And 958 with big endian |
| [11:39:28] | <mstenta[m]> | Bizarre. I don't know what would cause that. Could it be some lower level database configuration? |
| [11:40:15] | <symbioquine[m]> | Yeah, or even maybe some custom tooling of mine... I'm just surprised that so much of the data is that way. |
| [11:40:30] | <symbioquine[m]> | I would have expected more of a mix. |
| [11:41:39] | <mstenta[m]> | I've never dealt with this before. Is it possible to run an update query to convert them? |
| [11:42:13] | <mstenta[m]> | Oh could it be related to WFS operations from QGIS? |
| [11:44:01] | <symbioquine[m]> | Maybe, but it seems unlikely. As I recall, most of my geometries I mapped via my earlier WFS proxy which used the farmOS API |
| [11:45:30] | <symbioquine[m]> | In case it's helpful to other folks, the query I'm using to see the big endian geometries in my 1.x (MariaDB) DB is `SELECT entity_type, entity_id, SUBSTR(HEX(field_farm_geofield_geom), 1, 20) FROM `field_data_field_farm_geofield` WHERE SUBSTR(HEX(field_farm_geofield_geom), 1, 2) = '01'; ` |
| [11:49:23] | <symbioquine[m]> | symbioquine[m]: or rather that one shows the little endian ones :) it's `= '00'` at the end to show the big endian ones |
| [11:49:53] | <symbioquine[m]> | <mstenta[m]> "I've never dealt with this..." <- I'd imagine :) |
| [11:50:05] | <symbioquine[m]> | I guess I'll have to figure that out |
| [12:20:05] | <symbioquine[m]> | <mstenta[m]> "And the big endian geometries..." <- I think so |
| [12:23:47] | <symbioquine[m]> | Yeah, they all seem to show up correctly in 1.x |
| [12:24:22] | <mstenta[m]> | interesting - well maybe the geofield code doesn't care either way |
| [12:24:23] | <symbioquine[m]> | I'm guessing it's delegating to the geos library which maybe does support both endian-nesses |
| [12:24:26] | <mstenta[m]> | but geophp does |
| [12:24:49] | <mstenta[m]> | oh maybe geos plays a role too yea |
| [12:24:53] | <mstenta[m]> | lots of layers :-) |
| [12:24:58] | <symbioquine[m]> | yeah |
| [12:25:01] | <mstenta[m]> | geos / geophp / geofield |
| [12:32:17] | <symbioquine[m]> | I think in 1.x it may be loading via https://git.drupalcode.org/project/geofield/-/blob/7.x-2.x/geofield.modu... but I don't see the equivalent code path (to delegate to geos) in the 8.x-1.x version of geofield. |
| [12:34:27] | <symbioquine[m]> | Which I think could mean that data encoded/written on a big endian system via geos, (even with the vanilla farmOS docker container) might not migrate/display correctly in 2.x. |
| [12:36:42] | <symbioquine[m]> | That said, I think that should be a pretty rare scenario. |
| [12:37:23] | <mstenta[m]> | Ok yea I'm curious what might have caused big-endian usage in your db |
| [12:38:11] | <mstenta[m]> | <symbioquine[m]> "Maybe, but it seems unlikely. As..." <- was the WFS proxy writing binary data directly to that db column? |
| [12:38:35] | <symbioquine[m]> | I think I must have written a lot of the data via the python package `geomet` which seems to be writing the data as big-endian. |
| [12:38:44] | <mstenta[m]> | ahh ok |
| [12:38:49] | <symbioquine[m]> | mstenta[m]: The proxy used the farmOS APIs |
| [12:38:54] | <mstenta[m]> | oh ok |
| [12:39:44] | <symbioquine[m]> | The more recent WFS contrib module of mine uses the php entity API so that should be okay too. |
| [16:09:55] | <symbioquine[m]> | FYI: https://farmos.discourse.group/t/rewriting-farmos-1-x-geometry-wkb-to-be... |
