commit bf1b4130a469463040a3974a6b58a3e43149de5f
parent 208db0bba699fd529d4b7bc871dfa5e2ccc54b2c
Author: Antoine Amarilli <a3nm@a3nm.net>
Date: Mon, 6 Jan 2014 01:51:37 +0100
Merge branch 'master' of gitorious.org:drime/drime
Diffstat:
4 files changed, 44 insertions(+), 12 deletions(-)
diff --git a/README b/README
@@ -22,6 +22,16 @@ folder, and also "frhyme.py" (see plint's documentation).
drime also requires PyMySQL <http://www.pymysql.org> and Flask
<http://flask.pocoo.org/> for Python3.
+drime requires a working MySQL installation (package mysql-server on Debian
+systems) and the instructions in this README require the MySQL client (package
+mysql-client) to populate the database. You will probably want to create a
+database and a user for drime, probably along the lines of the following SQL
+commands executed using "mysql -u root -p":
+
+ CREATE DATABASE drime;
+ CREATE USER 'drime'@'localhost' IDENTIFIED BY 'mysecretpassword';
+ GRANT ALL ON drime.* TO 'drime'@'localhost';
+
== 3. Generating the DB ==
The program database isn't shipped, but scripts are provided to build it
@@ -41,7 +51,9 @@ If this is not your first attempt, unzip might prompt you about overwriting
existing files. Answer 'A'.
To import the output of lexique2sql.sh in a MySQL database (on localhost,
-database 'drime', as user 'drime', interactive password authentication), run:
+database 'drime', as user 'drime', interactive password authentication, the
+password being "mysecretpassword" if you followed the instructions of the
+previous section), run:
cat output.sql |
cat <(echo 'use drime;') - |
@@ -50,7 +62,7 @@ database 'drime', as user 'drime', interactive password authentication), run:
mysql --default-character-set=utf8 -D drime -u drime -p
To monitor progress, use the pv utility (replace "cat output.sql" by
-"pv -l output.sql".
+"pv -l output.sql").
== 4. Using the DB ==
@@ -82,9 +94,17 @@ specifics are left as an unpleasant exercise to the reader.
Run:
- ./query.py QUERY [NSYL [GENDER]]
+ ./query.py QUERY [NSYL [GENDER [CLASSICAL [PAGE]]]]
+
+The parameters are the same as the four fields of the web interface, with the
+option to specify a page. For instance, to request rhymes for "chanter" with
+between 2-4 syllables, respecting genre but not classical constraints, run:
+
+ ./query.py chanter 2-4 true true
+
+To get the second page of results, run:
-The parameters are the same as the three fields of the web interface.
+ ./query.py chanter 2-4 true true 2
Results are returned in a raw format (if the query is successful). To
pretty-print, do:
diff --git a/TODO b/TODO
@@ -1,5 +1,7 @@
-- disable classical
+- disable classical correctly on web interface
- only use python3
- change db module
- refactor common.py with plint
-- lien pour forcer la langue
+- lien pour forcer la langue sur l'interface web
+- meilleure notion de la rime
+- approche automatique, à base d'un corpus, avec exemples ?
diff --git a/query.py b/query.py
@@ -196,11 +196,11 @@ def do_query(word, phon, minsyll, maxsyll, elide, gender, classical, offset, siz
row['phon'] = to_xsampa(row['phon'])
result2.append(row)
c += 1
- if c > size:
+ if c > offset + size:
break
seen.add(row['word'])
count = len(result)
- result2 = result2[:size]
+ result2 = result2[offset:offset+size]
#cursor = run_query('''
#SELECT count(t2.word)
@@ -211,13 +211,18 @@ def do_query(word, phon, minsyll, maxsyll, elide, gender, classical, offset, siz
if __name__ == '__main__':
def usage():
- print ("Usage: %s QUERY [NSYL [GENDER [CLASSICAL]]]" % sys.argv[0])
+ print ("Usage: %s QUERY [NSYL [GENDER [CLASSICAL [PAGE]]]]" % sys.argv[0])
try:
for p in [3, 4]:
if p < len(sys.argv):
sys.argv[p] = convert(sys.argv[p])
+ try:
+ # adjust page to number starting at 0
+ sys.argv[5] = max(int(sys.argv[5]) - 1, 0)
+ except IndexError:
+ pass # page not specified
r, c, sure = query(*sys.argv[1:])
- except BadValues:
+ except (BadValues, ValueError, TypeError):
print ("Bad values passed as arguments.")
usage()
sys.exit(4)
diff --git a/templates/about.html b/templates/about.html
@@ -68,6 +68,8 @@ résultats dérivés d'un autre résultat mieux classé.</li>
<li>Le système ne produira pas de résultats suivant la règle classique de la
liaison supposée.</li>
<li>Le système ne renvoie que les {{ pagesize }} meilleurs résulats.</li>
+<li>Le système peut faire une interprétation trop stricte des règles classiques
+(ça ne devrait pas arriver souvent).</li>
</ul>
<h2>Quels sont les résultats surnuméraires ? (faux positifs)</h2>
@@ -78,7 +80,8 @@ faire une rime acceptable (par exemple "reprendre" et "prendre").</li>
<li>Le système autorisera des rimes pour l'œil abusives (vers la fin des
résultats).</li>
<li>Le système autorisera des rimes qui ne sont pas acceptables pour les règles
-classiques. C'est un problème si vous essayez de les suivre.</li>
+classiques, parce qu'il est trop tolérant quant à l'interprétation des règles
+classiques.</li>
</ul>
@@ -162,6 +165,8 @@ derived from a previous result.</li>
<li>The system will not accept rhymes for the classical rule of <em>liason
supposée</em>.</li>
<li>The system will only return the {{ pagesize }} best results.</li>
+<li>The system may be too strict when enforcing classical rules (it shouldn't
+happen very often, though).</li>
</ul>
<h2>Which results are unnecessary? (false positives)</h2>
@@ -171,7 +176,7 @@ make a decent rhyme with it (eg. "reprendre" for "prendre").</li>
<li>The system will provide results with a wrong syllable count.</li>
<li>The system will provide abusive visual rhymes (at the end of the results).</li>
<li>The system will provide rhymes which do not follow classical constraints,
-which is a problem if you want to follow them.</li>
+because its interpretation of them may be too liberal.</li>
</ul>
<h2>How are results sorted?</h2>
