How to parse Sumerian texts

The Sumerian texts are lemmatized, and we use the information from the lemmas to help with the tagging. The hope is that the result is so accurate that a correction stage can be dispensed with.

replace bad line breaks from Word

cat $input | tr '\015' '\012' > $output

protect CODE material (line numbers, comments, and the like) by enclosing it in angle brackets and tagging as CODE
```
enclose-code-material $input > $output
```
replace spaces inside CODE material with underbars
```
fix-bad-spaces $input > $output
```
put #lem line on same line as word forms (in emacs)
align lemmas with words in short lines (<= 5 words/lemmas)
```
align-lemmas $input > $output
```
prepare remaining long lines
```
prepare-long-lines $input > $output
```
align lemmas with words with one word and its lemma per line
find spaces in lemmas and replace by underbars
```
grep '<[^>]*[ ]+[^>]*>/LEMMA' $input
```

convert spaces to newlines

cat $input | tr  -s ' ' '\012' > $output

sanity checks

grep '>' $input | grep -v '<'
grep '<' $input | grep -v '>'
grep LEMMA $input | grep -v '<'
grep LEMMA $input | grep -v '>'
grep LEMMA $input | grep ' '
grep 'LEMMA.' $input
grep 'LEMMA.*LEMMA' $input
grep '/.*/' $input

put lemmas on same line as their words (in emacs); not all lines will have a lemma
construct tagging scripts for open-class items from lists of lemmas
- tag-adjectives
- tag-adverbs
- tag-nouns
- tag-verbs
tag text with above scripts and tag-other
split off case markers
```
fix-case-markers $input > $output
```
check for remaining case markers
```
grep ']/[a-]z>/' $file
```
deal with verb complex - NEEDS FURTHER WORK
if not yet done, divide sentence tokens
if using blanklines to divide sentence tokens, replace blanklines by BLANKLINE/CODE, EOS/CODE, or the like (blanklines are gobbled up by the following script)
sanity check - does each line have exactly one slash?

replace special characters (parens, slash)

protect-special-characters $input > $output

convert from word/tag to (tag word) format
```
pos-to-psd2 $input > $output
```
restore blanklines between sentence tokens
add wrapper parens and root node (S-MAT) (in emacs)
add syntactic structure
```
PARS1
```