source: modelado_topicos/topic_explorer/views.py @ fb25bf8

from django.shortcuts import render

from django.conf import settings

from django.http import HttpResponse, HttpResponseRedirect, HttpResponseServerError, StreamingHttpResponse

import json
import colorlib
import itertools
from vsm.corpus import Corpus
from vsm.model.ldacgsmulti import LdaCgsMulti as LCM
from vsm.viewer.ldagibbsviewer import LDAGibbsViewer as LDAViewer
from vsm.viewer.wrappers import doc_label_name



#path = settings.PATH 
corpus_file = settings.CORPUS_FILE
context_type = settings.CONTEXT_TYPE
model_pattern = settings.MODEL_PATTERN
topics = settings.TOPICS
corpus_name = settings.CORPUS_NAME
icons = settings.ICONS

corpus_link = settings.CORPUS_LINK
topics_range = [int(item) for item in settings.TOPICS.split(',')]
doc_title_format = settings.DOC_TITTLE_FORMAT
doc_url_format = settings.DOC_URL_FORMAT

#global lda_m, lda_v


lda_c = Corpus.load(corpus_file)
#lda_m = LCM.load(model_pattern.format(k))
#lda_v = LDAViewer(lda_c, lda_m)
label = lambda x: x

def dump_exception():
    import sys,traceback
    exc_type, exc_value, exc_traceback = sys.exc_info()
    print "*** print_tb:"
    traceback.print_tb(exc_traceback, limit=1, file=sys.stdout)
    print "*** print_exception:"
    traceback.print_exception(exc_type, exc_value, exc_traceback, limit=2, file=sys.stdout)
    return HttpResponseServerError(str(exc_value))


def doc_topic_csv(request, doc_id):
    data = lda_v.doc_topics(doc_id)

    output=StringIO()
    writer = csv.writer(output)
    writer.writerow(['topic','prob'])
    writer.writerows([(t, "%6f" % p) for t,p in data])

    return HttpResponse(output.getvalue())

def doc_csv(request, doc_id, threshold=0.2):
    data = lda_v.sim_doc_doc(doc_id)

    output=StringIO()
    writer = csv.writer(output)
    writer.writerow(['doc','prob'])
    writer.writerows([(d, "%6f" % p) for d,p in data if p > threshold])

    return HttpResponse(output.getvalue())

def topic_json(request,k_param,topic_no, N=40):
    global lda_v
    lda_m = LCM.load(model_pattern.format(k_param))
    lda_v = LDAViewer(lda_c, lda_m)
    try:
        N = int(request.query.n)
    except:
        pass

    if N > 0:
        data = lda_v.dist_top_doc([int(topic_no)])[:N]
    else:
        data = lda_v.dist_top_doc([int(topic_no)])[N:]
        data = reversed(data)

    docs = [doc for doc,prob in data]
    doc_topics_mat = lda_v.doc_topics(docs)

    js = []
    for doc_prob, topics in zip(data, doc_topics_mat):
        doc, prob = doc_prob
        js.append({'doc' : doc, 'label': label(doc), 'prob' : 1-prob,
            'topics' : dict([(str(t), p) for t,p in topics])})

    return HttpResponse(json.dumps(js))

def doc_topics(request,doc_id, N=40):
    try:
        try:
            N = int(request.query.n)
        except:
            pass
        if N > 0:
            data = lda_v.dist_doc_doc(doc_id)[:N]
        else:
            data = lda_v.dist_doc_doc(doc_id)[N:]
            data = reversed(data)

        docs = [doc for doc,prob in data]
        doc_topics_mat = lda_v.doc_topics(docs)

        js = []
        for doc_prob, topics in zip(data, doc_topics_mat):
            doc, prob = doc_prob
            js.append({'doc' : doc, 'label': label(doc), 'prob' : 1-prob,
                'topics' : dict([(str(t), p) for t,p in topics])})

        return HttpResponse(json.dumps(js))
    except:
        return dump_exception()

def topics(request):
    try:
        # populate entropy values
        data = lda_v.topic_oscillations()

        colors = [itertools.cycle(cs) for cs in zip(*colorlib.brew(3,n_cls=4))]
        factor = len(data) / len(colors)

        js = {}
        for rank,topic_H in enumerate(data):
            topic, H = topic_H
            js[str(topic)] = {
                "H" : H,
                "color" : colors[min(rank / factor, len(colors)-1)].next()
            }

        # populate word values
        data = lda_v.topics()
        for i,topic in enumerate(data):
            js[str(i)].update({'words' : dict([(w, p) for w,p in topic[:10]])})

        return HttpResponse(json.dumps(js))
    except:
        return dump_exception()

def docs(request):
    try:
        docs = lda_v.corpus.view_metadata(context_type)[doc_label_name(context_type)]
        js = list()
        for doc in docs:
            js.append({
                'id': doc,
                'label' : label(doc)
            })

        return HttpResponse(json.dumps(js))
    except:
        return dump_exception()

def index(request):
    global lda_m,lda_v
    lda_m = LCM.load(model_pattern.format(10))
    lda_v = LDAViewer(lda_c, lda_m)
    template = 'index.html'
    return render(request,template,
        {'filename':None,
         'corpus_name' : corpus_name,
         'corpus_link' : corpus_link,
         'context_type' : context_type,
         'topics_range' : topics_range,
         'doc_title_format' : doc_title_format,
         'doc_url_format' : doc_url_format})

def visualize(request,k_param,filename=None,topic_no=None):
    global lda_m,lda_v
    lda_m = LCM.load(model_pattern.format(k_param))
    lda_v = LDAViewer(lda_c, lda_m)
    template = 'index.html'
    return render(request,template,
        {'filename':filename,
         'k_param':k_param,
         'topic_no':topic_no,
         'corpus_name' : corpus_name,
         'corpus_link' : corpus_link,
         'context_type' : context_type,
         'topics_range' : topics_range,
         'doc_title_format' : doc_title_format,
         'doc_url_format' : doc_url_format})