link-grammar
link-grammar copied to clipboard
opencog
The CMU Link Grammar natural language parser
(EDIT: Since I have already written it in the wrong box, I leave it here...) It seems there are several problems in this code. Also, there are several other problems...
Currently, this line: https://github.com/opencog/link-grammar/blob/af7033e541160a0d0b941d9ed70c7dff7f979361/link-parser/lg_readline.c#L375 causes the `.lg_history` file to be written in the current working directory, resulting in LG litter everywhere. It's getting annoying. It should be changed to either...
I have an idea of how to memory-spare parse choice elements, just like parse set elements are shared through the x-table. Currently, they cannot be shared because each has a...
After experimentation, this seems to ... build without error. No idea if the result is usable, or not.
If `form_match_list()` would return a match list that more than 4090 of its elements would have a linkage (for both LHS and RHS), then a segfault would happen. As I...
It can get very large during MST parsing, and just takes up RAM that extract-linkages could use. So I'd like to free it, but I cannot, because parsing with null...
Heh. > @MXs actually mean "connect to one or more `MXs` Yes it does. I was wrong; I misremembered. I was so used to seeing `{@MXs}` that I forgot the...
Per discussions in #1403, pertaining to this figure, which shows `pool_size(sent->Table_tracon_pool)` in `count.c`  Each horizontal line is a re-alloc of 16K elements. Counting the line, I see that there...
In recent CI checks, in the Emscripten configuration LT_INIT checks for `threads.h`, as can be seen from the following output (still exists in PR #1365 from Jan 13 2023): ```...
Github seems to be using two different sets of autoconf macros in different instances. The one on this git repo is "sane" and seems to be the same as what...