def lsiformatoutput(findshtml: str, workssearched: int, matches: list,
searchobject: SearchObject) -> str:
"""
should use OutputObject() instead
:param findshtml:
:param workssearched:
:param searchobject:
:param activepoll:
:param starttime:
:return:
"""
so = searchobject
activepoll = so.poll
output = SearchOutputObject(so)
output.found = findshtml
output.js = insertbrowserclickjs('browser')
output.setscope(workssearched)
output.title = 'Sentences that are reminiscent of »{skg}«'.format(
skg=so.seeking)
output.thesearch = output.title
output.htmlsearch = 'sentences that are reminiscent of <span class="sought">»{skg}«</span>'.format(
skg=so.seeking)
output.resultcount = '{n} sentences above the cutoff'.format(
n=len(matches))
output.searchtime = so.getelapsedtime()
activepoll.deactivate()
jsonoutput = json.dumps(output.generateoutput())
return jsonoutput
def threejsgraphofvectors(sentencetuples, workssearched, so, vectorspace):
"""
unused parameters so that the shape of this function's inputs can match other parallel functions
see https://github.com/tobydoig/3dword2vec
:param sentencetuples:
:param workssearched:
:param so:
:param vectorspace:
:return:
"""
output = SearchOutputObject(so)
graphdata = reducetothreedimensions(so, vectorspace)
output.found = str()
output.htmlsearch = str()
output.found = trheedimensionalhtml()
output.js = threedimensionaljs(graphdata)
jsonoutput = json.dumps(output.generateoutput())
# print('jsonoutput', jsonoutput)
return jsonoutput
def ldatopicsgenerateoutput(ldavishtmlandjs: str, searchobject: SearchObject):
"""
pyLDAvis.prepared_data_to_html() outputs something that is almost pure JS and looks like this:
<link rel="stylesheet" type="text/css" href="https://cdn.rawgit.com/bmabey/pyLDAvis/files/ldavis.v1.0.0.css">
<div id="ldavis_el7428760626948328485476648"></div>
<script type="text/javascript">
var ldavis_el7428760626948328485476648_data = {"mdsDat": ...
}
</script>
instance = {
'maxfeatures': 2000,
'components': 15, # topics
'maxfreq': .75, # fewer than n% of sentences should have this word (i.e., purge common words)
'minfreq': 5, # word must be found >n times
'iterations': 12,
'mustbelongerthan': 3
}
:param ldavishtmlandjs:
:param workssearched:
:param settings:
:param searchobject:
:return:
"""
so = searchobject
activepoll = so.poll
output = SearchOutputObject(so)
workssearched = len(so.searchlist)
vv = searchobject.vectorvalues
settings = {
'maxfeatures': vv.ldamaxfeatures,
'components': vv.ldacomponents, # topics
'maxfreq': vv.
ldamaxfreq, # fewer than n% of sentences should have this word (i.e., purge common words)
'minfreq': vv.ldaminfreq, # word must be found >n times
'iterations': vv.ldaiterations,
'mustbelongerthan': vv.ldamustbelongerthan
}
lines = ldavishtmlandjs.split('\n')
lines = [re.sub(r'\t', str(), l) for l in lines if l]
lines.reverse()
thisline = str()
html = list()
while not re.search(r'<script type="text/javascript">', thisline):
html.append(thisline)
try:
thisline = lines.pop()
except IndexError:
# oops, we never found the script...
thisline = '<script type="text/javascript">'
# we cut '<script>'; now drop '</script>'
lines.reverse()
js = lines[:-1]
findshtml = '\n'.join(html)
findsjs = '\n'.join(js)
ldacssurl = r'https://cdn.rawgit.com/bmabey/pyLDAvis/files/ldavis.v1.0.0.css'
ldacsslocal = '/css/ldavis.css'
findshtml = re.sub(ldacssurl, ldacsslocal, findshtml)
# brittle: ldavis might change its URLs between versions, etc.
# should probably make this conditional upon the presence of the file locally...
ldajsurl = r'https://cdn.rawgit.com/bmabey/pyLDAvis/files/ldavis.v1.0.0.js'
ldajslocal = '/static/jsforldavis.js'
findsjs = re.sub(ldajsurl, ldajslocal, findsjs)
# this next will break the reloaded figure: hm...
# d3jsurl = r'https://cdnjs.cloudflare.com/ajax/libs/d3/3.5.5/d3.min'
# d3jslocal = '/static/jsd3'
# findsjs = re.sub(d3jsurl, d3jslocal, findsjs)
#
# print('findsjs',findsjs)
who = str()
where = '{n} authors'.format(n=searchobject.numberofauthorssearched())
if searchobject.numberofauthorssearched() == 1:
a = authordict[searchobject.searchlist[0][:6]]
who = a.akaname
where = who
if workssearched == 1:
try:
w = workdict[searchobject.searchlist[0]]
w = w.title
except KeyError:
w = str()
where = '{a}, <worktitle>{w}</worktitle>'.format(a=who, w=w)
output.title = 'Latent Dirichlet Allocation'
output.found = findshtml
output.js = findsjs
output.setscope(workssearched)
output.sortby = 'weight'
output.thesearch = 'thesearch'.format(skg='')
output.resultcount = 'the following topics'
output.htmlsearch = '{n} topics in {w}'.format(n=settings['components'],
w=where)
output.searchtime = so.getelapsedtime()
activepoll.deactivate()
jsonoutput = json.dumps(output.generateoutput())
return jsonoutput
def nearestneighborgenerateoutput(findshtml: str, mostsimilar: list,
imagename: str, workssearched: int,
searchobject: SearchObject) -> str:
"""
:param findshtml:
:param mostsimilar:
:param imagename:
:param workssearched:
:param searchobject:
:param activepoll:
:param starttime:
:return:
"""
vectorsearchwaslemmatized = True
so = searchobject
activepoll = so.poll
output = SearchOutputObject(so)
output.image = imagename
findsjs = generatevectorjs()
try:
lm = so.lemma.dictionaryentry
except AttributeError:
# AttributeError: 'NoneType' object has no attribute 'dictionaryentry'
vectorsearchwaslemmatized = False
lm = so.seeking
try:
pr = so.proximatelemma.dictionaryentry
except AttributeError:
# proximatelemma is None
pr = None
if vectorsearchwaslemmatized:
extrastringone = 'all forms of '
ht = 'all {n} known forms of <span class="sought">»{skg}«</span>'.format(
n=len(so.lemma.formlist), skg=lm)
else:
extrastringone = str()
ht = '<span class="sought">»{skg}«</span>'.format(skg=lm)
output.title = 'Neighbors for {es}»{skg}«'.format(skg=lm,
pr=pr,
es=extrastringone)
output.found = findshtml
output.js = findsjs
try:
output.setresultcount(len(mostsimilar), 'proximate terms to graph')
except TypeError:
pass
output.setscope(workssearched)
output.thesearch = '{es}»{skg}«'.format(skg=lm, es=extrastringone)
output.htmlsearch = ht
output.sortby = 'proximity'
output.image = imagename
output.searchtime = so.getelapsedtime()
jsonoutput = json.dumps(output.generateoutput())
activepoll.deactivate()
if isinstance(activepoll, RedisProgressPoll):
activepoll.deleteredispoll()
del activepoll
return jsonoutput
def executesearch(searchid: str, so=None, req=request) -> JSON_STR:
"""
the interface to all of the other search functions
tell me what you are looking for and i'll try to find it
the results are returned in a json bundle that will be used to update the html on the page
note that cosdistbysentence vector queries also flow through here: they need a hitdict
overview:
buildsearchobject() and then start modifying elements of the SearchObject
build a search list via compilesearchlist()
modify search list via flagexclusions()
modify search list via calculatewholeauthorsearches()
build search list restrictions via indexrestrictions()
search via searchdispatcher()
format results via buildresultobjects()
:return:
"""
pollid = validatepollid(searchid)
if not so:
# there is a so if singlewordsearch() sent you here
probeforsessionvariables()
so = buildsearchobject(pollid, req, session)
frozensession = so.session
progresspolldict[pollid] = ProgressPoll(pollid)
so.poll = progresspolldict[pollid]
so.poll.activate()
so.poll.statusis('Preparing to search')
nosearch = True
output = SearchOutputObject(so)
allcorpora = [
'greekcorpus', 'latincorpus', 'papyruscorpus', 'inscriptioncorpus',
'christiancorpus'
]
activecorpora = [c for c in allcorpora if frozensession[c]]
if (len(so.seeking) > 0 or so.lemma or frozensession['tensorflowgraph']
or frozensession['topicmodel']) and activecorpora:
so.poll.statusis('Compiling the list of works to search')
so.searchlist = compilesearchlist(listmapper, frozensession)
if so.searchlist:
# do this before updatesearchlistandsearchobject() which collapses items and cuts your total
workssearched = len(so.searchlist)
# calculatewholeauthorsearches() + configurewhereclausedata()
so = updatesearchlistandsearchobject(so)
nosearch = False
skg = None
prx = None
isgreek = re.compile(
'[α-ωϲἀἁἂἃἄἅἆἇᾀᾁᾂᾃᾄᾅᾆᾇᾲᾳᾴᾶᾷᾰᾱὰάἐἑἒἓἔἕὲέἰἱἲἳἴἵἶἷὶίῐῑῒΐῖῗὀὁὂὃὄὅόὸὐὑὒὓὔὕὖὗϋῠῡῢΰῦῧύὺᾐᾑᾒᾓᾔᾕᾖᾗῂῃῄῆῇἤἢἥἣὴήἠἡἦἧὠὡὢὣὤὥὦὧᾠᾡᾢᾣᾤᾥᾦᾧῲῳῴῶῷώὼ]'
)
if so.lemmaone:
so.termone = wordlistintoregex(so.lemma.formlist)
skg = so.termone
if re.search(isgreek, skg):
# 'v' is a problem because the lemmata list is going to send 'u'
# but the greek lemmata are accented
so.usecolumn = 'accented_line'
if so.lemmatwo:
so.termtwo = wordlistintoregex(so.lemmatwo.formlist)
prx = so.termtwo
if re.search(isgreek, prx):
so.usecolumn = 'accented_line'
so.setsearchtype()
thesearch = so.generatesearchdescription()
htmlsearch = so.generatehtmlsearchdescription()
# now that the SearchObject is built, do the search...
hits = precomposedsqlsearch(so)
so.poll.statusis('Putting the results in context')
# hits is List[dbWorkLine]
hitdict = sortresultslist(hits, so, authordict, workdict)
if so.vectorquerytype == 'cosdistbylineorword':
# print('executesearch(): h - cosdistbylineorword')
# take these hits and head on over to the vector worker
output = findabsolutevectorsfromhits(so, hitdict, workssearched)
del progresspolldict[pollid]
return output
resultlist = buildresultobjects(hitdict, authordict, workdict, so)
so.poll.statusis('Converting results to HTML')
sandp = rewriteskgandprx(skg, prx, htmlsearch, so)
skg = sandp['skg']
prx = sandp['prx']
htmlsearch = sandp['html']
for r in resultlist:
r.lineobjects = flagsearchterms(r, skg, prx, so)
if so.context > 0:
findshtml = htmlifysearchfinds(resultlist, so)
else:
findshtml = nocontexthtmlifysearchfinds(resultlist)
if hipparchia.config['INSISTUPONSTANDARDANGLEBRACKETS']:
findshtml = gtltsubstitutes(findshtml)
findsjs = insertbrowserclickjs('browser')
resultcount = len(resultlist)
if resultcount < so.cap:
hitmax = False
else:
hitmax = True
output.title = thesearch
output.found = findshtml
output.js = findsjs
output.setresultcount(resultcount, 'passages')
output.setscope(workssearched)
output.searchtime = so.getelapsedtime()
output.thesearch = thesearch
output.htmlsearch = htmlsearch
output.hitmax = hitmax
if nosearch:
if not activecorpora:
output.reasons.append('there are no active databases')
if len(so.seeking) == 0:
output.reasons.append('there is no search term')
if len(so.seeking) > 0 and len(so.searchlist) == 0:
output.reasons.append('zero works match the search criteria')
output.title = '(empty query)'
output.setresultcount(0, 'passages')
output.explainemptysearch()
so.poll.deactivate()
jsonoutput = json.dumps(output.generateoutput())
del progresspolldict[pollid]
return jsonoutput
def generateabsolutevectorsoutput(listsofwords: list, workssearched: list,
searchobject, vtype: str):
"""
:return:
"""
so = searchobject
vv = so.vectorvalues
activepoll = so.poll
# find all words in use
allwords = findsetofallwords(listsofwords)
# print('allwords', allwords)
# find all possible forms of all the words we used
# consider subtracting some set like: rarewordsthatpretendtobecommon = {}
activepoll.statusis('Finding headwords')
morphdict = getrequiredmorphobjects(allwords, furtherdeabbreviate=True)
morphdict = convertmophdicttodict(morphdict)
# find all possible headwords of all of the forms in use
# note that we will not know what we did not know: count unparsed words too and deliver that as info at the end?
allheadwords = dict()
for m in morphdict.keys():
for h in morphdict[m]:
allheadwords[h] = m
if so.lemma:
# set to none for now
subtractterm = None
else:
subtractterm = so.seeking
activepoll.statusis('Building vectors')
vectorspace = buildrudimentaryvectorspace(allheadwords,
morphdict,
listsofwords,
subtractterm=subtractterm)
# for k in vectorspace.keys():
# print(k, vectorspace[k])
if so.lemma:
focus = so.lemma.dictionaryentry
else:
focus = so.seeking
activepoll.statusis('Calculating cosine distances')
cosinevalues = caclulatecosinevalues(focus, vectorspace,
allheadwords.keys())
# cosinevalues = vectorcosinedispatching(focus, vectorspace, allheadwords.keys())
# print('generatevectoroutput cosinevalues', cosinevalues)
# apply the threshold and drop the 'None' items
threshold = 1.0 - vv.localcutoffdistance
falseidentity = .02
cosinevalues = {
c: 1 - cosinevalues[c]
for c in cosinevalues
if cosinevalues[c] and falseidentity < cosinevalues[c] < threshold
}
mostsimilar = [(c, cosinevalues[c]) for c in cosinevalues]
mostsimilar = sorted(mostsimilar, key=lambda t: t[1], reverse=True)
findshtml = formatnnmatches(mostsimilar, vv)
# next we look for the interrelationships of the words that are above the threshold
activepoll.statusis('Calculating metacosine distances')
imagename = graphbliteraldistancematches(focus, mostsimilar, so)
findsjs = generatevectorjs()
output = SearchOutputObject(so)
output.title = 'Cosine distances to »{skg}«'.format(skg=focus)
output.found = findshtml
output.js = findsjs
if not so.session['cosdistbylineorword']:
space = 'related terms in {s} {t}'.format(s=len(listsofwords), t=vtype)
else:
dist = so.session['proximity']
scale = {'words': 'word', 'lines': 'line'}
if int(dist) > 1:
plural = 's'
else:
plural = str()
space = 'related terms within {a} {b}{s}'.format(
a=dist, b=scale[so.session['searchscope']], s=plural)
found = max(vv.neighborscap, len(cosinevalues))
output.setresultcount(found, space)
output.setscope(workssearched)
if so.lemma:
xtra = 'all forms of '
else:
xtra = str()
output.thesearch = '{x}»{skg}«'.format(x=xtra, skg=focus)
output.htmlsearch = '{x}<span class="sought">»{skg}«</span>'.format(
x=xtra, skg=focus)
output.sortby = 'distance with a cutoff of {c}'.format(
c=vv.localcutoffdistance)
output.image = imagename
output.searchtime = so.getelapsedtime()
activepoll.deactivate()
jsonoutput = json.dumps(output.generateoutput())
return jsonoutput
def tsnegraphofvectors(sentencetuples, workssearched, so, vectorspace):
"""
lifted from https://radimrehurek.com/gensim/auto_examples/tutorials/run_word2vec.html#sphx-glr-auto-examples-tutorials-run-word2vec-py
unused parameters so that the shape of this function's inputs can match other parallel functions
:param sentencetuples:
:param workssearched:
:param so:
:param vectorspace:
:return:
"""
plotdict = reducetotwodimensions(vectorspace)
xvalues = plotdict['xvalues']
yvalues = plotdict['yvalues']
labels = plotdict['labels']
# random.seed(0)
plt.figure(figsize=(12, 12))
# https://jonasjacek.github.io/colors/
plt.scatter(xvalues, yvalues, color='#c6c6c6')
# Label randomly subsampled 25 data points
#
indices = list(range(len(labels)))
selected_indices = random.sample(indices, 25)
for i in selected_indices:
plt.annotate(labels[i], (xvalues[i], yvalues[i]))
graphobject = BytesIO()
plt.savefig(graphobject)
plt.clf()
plt.close()
graphobject = graphobject.getvalue()
imagename = storevectorgraph(graphobject)
# print('http://localhost:5000/getstoredfigure/{i}'.format(i=imagename))
output = SearchOutputObject(so)
output.image = imagename
findsjs = generatevectorjs()
htmltemplate = """
<p id="imagearea"></p>
"""
output.found = str()
output.htmlsearch = str()
output.found = htmltemplate
output.js = findsjs
jsonoutput = json.dumps(output.generateoutput())
# print('jsonoutput', jsonoutput)
return jsonoutput