We have figured out a way to fix this problem but have not yet updated the code accordingly.- Two languages added to the Collatinus registry: Dutch and Euskara. One of the main problems we encounter when punctuating the LASLA database with a text from PHI is that quite often sentences will be in a different order. The v_reading module contains a few classes that are not all that related but it does contain the key class where we put all of our analysis together and produce a text complete with vowel macronization, accent marks and elision. Presently roughly 90% of the lemmas in LASLA have been matched to lemmas in Collatinus. In the j_lasla module we normalize the LASLA database and in the j_lasla2 we merge the database with the Collatinus database. I have managed to separate a lot of words into their constituent morphemes but much more work remains to be done. It turns out that the Oxford Latin Dictionary has already done a remarkable job with this. The m_stems module was meant to divide words into their constituent morphemes.
This work is done in the o_memorize module. This problem has mostly been solved but I still have not fully integrated all of the changes to the third declension into the new database. The third declension turns out to differ widely in practice then as listed in the Collatinus database. In the future I plan to get the vowel length from the Pedecerto database firstly, Oxford Latin Dictionary secondly.Īnother major problem we are working on is the third declension. I have not yet corrected this mistake because I'm going to change how I get the vowel length anyway. Suppose this author was B but B disagrees with A regarding the vowel length of the present stem. But suppose only two authors listed the vowel length for the perfect stem of "figo" which is "fixo". I would then use the vowel length for the author most in agreement with the Pedecerto database, call him A. So for example, suppose all five authors listed the vowel length for the present stem of "figo". One major error that I have not yet corrected is that although dictionary authors agree roughly 97% of the time for those syllables which are long or short by nature, and 80% of the time they agree on the vowel quantity for syllables which are long by position, I made the unfortunate decision to not always get the vowel length from the same authors for the perfect stem than I did for the main stem. The Collatinus database contains the syllable length but not the vowel length. Our main ambition with the Collatinus database is to restore the vowel length of the words. Because each text has its quirks these texts mostly have to be examined on an individual basis but with each text the job of restoring them will become easier. So far this has been done for Petronius' Satyricon. This code has several ambitions but certainly one of the most important ambitions is to reconstruct the LASLA database into punctuated text complete with macrons, elision and accent marks. In this way, if a mistake has been made, it can be tracked down and more of its kind can be spotted. Because it is important to see how we get from the separate databases to the new database, the z_begin module will loop through the entire code from the beginning to the end. The Packhard Humanities Institute Corpus or PHI.No definitions or examples of word usage is available. We also use the vowel quantity and the known spelling variants of these words from the Oxford Latin Dictionary. Only about 75% of this database is currently being used.ĭue to copyright restrictions only a list of words found in the Oxford Latin Dictionary is available.
#Collatinus web software#
This software merges several Latin databases into one. The z_begin method takes roughly 20 minutes to complete.
As an example of this codes functionality we repunctuate the LASLA database of the Satyricon into a punctuated text with macrons, accents and elision.
#Collatinus web download#
These will both download files and also create new ones. To use this software, after performing a pip install, you must perform two steps:
0 kommentar(er)
