- 1be7b13 add SuperExpress with a known offset-related issue, for investigation; requires Xmx set to "8g" by Piotr Banski · 2 years, 3 months ago
- ad4b7ce modified to preserve the analogy to ginkgo by Piotr Banski · 2 years, 5 months ago
- 381e0c0 handle cases where no NE information is present in an NKJP NE document (don't produce named.xml then) by Piotr Banski · 2 years, 6 months ago
- b28e588 catalog fixed. script prepared for processing, morpho files have some new data now (from the new NKJP version) by Piotr Banski · 2 years, 6 months ago
- a78e59d new dataset, up to NE by Piotr Banski · 2 years, 6 months ago
- a44cd7a produce morpho.xml with additional information: translit as 'orig' and all the morphologically possible values by Piotr Banski · 2 years, 6 months ago
- d2b78b8 cleaned up, before adding more info to morpho by Piotr Banski · 2 years, 6 months ago
- ba6cc63 successfully switched to ann_morpho for morpho.xml; next step: cleanup + more info in morpho descriptions by Piotr Banski · 2 years, 6 months ago
- 43b9db0 successfully switched to ann_morpho for structure.xml, with some lingering questions asked via slack by Piotr Banski · 2 years, 6 months ago
- 081c5de successfully switched to ann_morpho for data.xml by Piotr Banski · 2 years, 6 months ago
- a0a9fc0 hopefully final refactoring before the switch by Piotr Banski · 2 years, 6 months ago
- faa910f more refactoring, for readability by Piotr Banski · 2 years, 6 months ago
- f959069 commiting the state after minor refactoring of parameter/variable names, still before eliminating ann_segmentation by Piotr Banski · 2 years, 6 months ago
- 763b41f recording the state before the transfer to use ann_morphosyntax as the basis (due to the manual corrections present there) by Piotr Banski · 2 years, 6 months ago
- 65a6d0b produce an initial version of named.xml, with just placeholders but also with properly computed offsets (walking ann_morphosyntactic) by Piotr Banski · 2 years, 6 months ago
- e1ac520 revert; the intended lookup from old ID to the new index across the entire tree would require a lot of magic and wouldn't be efficient at all by Piotr Banski · 2 years, 6 months ago
- 9397ca5 extend the indexing accumulator to be a mapping from NKJP index onto the accumulator-generated traverse value that gets turned into the KorAP index; this way we get free lookup of a sort, at least for cases where all we need is an old->new mapping by Piotr Banski · 2 years, 6 months ago
- ad3581f further optimization by Piotr Banski · 2 years, 6 months ago
- 8d2609a optimised tei:seg (total net time cut by half) by Piotr Banski · 2 years, 6 months ago
- c5950ce added placeholders for handling more layers of annotation; tei:seg in 'morpho' mode needs some streamlining (the profiler suggests) by Piotr Banski · 2 years, 6 months ago
- 92791a2 Eliminated spurious whitespace (which probably didn't hurt, but...) by Piotr Banski · 2 years, 6 months ago
- 1ae16bd this is the version used to derive the entire dataset; it doesn't need any parameters and is set to output comments in structure.xml, just in case we need them for debugging by Piotr Banski · 2 years, 6 months ago
- a51907c minimal goal reached: data should be now ingestible; there is a niche of inefficiency though that may prevent my desktop from processing the entire dataset (other than KOT) by Piotr Banski · 2 years, 6 months ago
- 09096ee many fixes, structure.xml all done, retaining comments with the surface forms for now by Piotr Banski · 2 years, 6 months ago
- fdc858a I think it's done. Will clean it tomorrow and extend to morpho and beyond. My goodness. by Piotr Banski · 2 years, 6 months ago
- 69f3c5f this is just for demonstration that the offsets are now done in the KorAP way by Piotr Banski · 2 years, 6 months ago
- 5fe4bae far from perfect, but the road is straight now; adding the structure.xml doc, a bit schizophrenic in the indexing by Piotr Banski · 2 years, 6 months ago
- 6a4a252 this version attempts to re-traverse the tree over 6k times per single output document with structure in it, and I can't seem to be able to help that. It does the necessary calculation perfectly, but, naturally, in doing so it crashes my desktop by Piotr Banski · 2 years, 6 months ago
- f8af3a9 just saving the next step by Piotr Banski · 2 years, 6 months ago
- 4f4c2d2 one step further and I just want to save it by Piotr Banski · 2 years, 7 months ago
- 9dc1000 begin the switch from text.xml to ann_segmentation.xml; for now, data.xml is properly created (whitespace and tokenization alternatives). A lot of code cleanup has not yet happened. by bansp · 2 years, 7 months ago
- d1bf1db before migration from calc_content_length to calc_offsets by bansp · 2 years, 8 months ago
- a8e5cf1 forgot to save, already pushed the previous one, sorry by bansp · 2 years, 8 months ago
- b599253 this is a safety commit, before I take some stuff apart by bansp · 2 years, 8 months ago
- f2b24e6 add ability to skip some document IDs as a comma-separated parameter by bansp · 2 years, 8 months ago
- e726b4a stylesheet redone for handling larger datasets; just struct and morpho for now, though by bansp · 2 years, 8 months ago
- 8f6700b initial modification that I need to commit by bansp · 2 years, 8 months ago
- 97ba7ce update the I5 DTD identifiers, just in case by bansp · 2 years, 8 months ago
- ba37fb9 update gitignore by bansp · 2 years, 8 months ago
- 9103aab attempt to add to the headers (they are black boxes) by bansp · 2 years, 9 months ago
- 3e5b20c fix structure.xml, create morpho.xml by bansp · 2 years, 9 months ago
- 5f84173 derive structure.xml; the script isn't optimized yet but I would like to submit the output for a check by bansp · 2 years, 9 months ago
- 608b102 fix textSigle by bansp · 2 years, 9 months ago
- 5e2d1c0 first touch: make sure that I can grab at the data and send it where I want it to go by bansp · 2 years, 9 months ago
- 68529a8 proof of concept that I'm reaching for the right info by bansp · 2 years, 9 months ago
- 0748682 correct empty namespace nodes in the output by bansp · 2 years, 9 months ago
- f79443e version working in oXygen, data.xml should be identical modulo whitespace; missing references: metadata.xml, text.rng by bansp · 2 years, 9 months ago
- 9a8ee3e Added initial conversion script and example xspec by Akron · 2 years, 10 months ago