def OnListCtrl1ListItemSelected(self, event):
#print "selected"
self.item_courant = event.m_itemIndex
#self.lig_donnees = donnees[self.item_courant]['nom']
db = getdata.query("select nom from Popu")
#print db.data[self.item_courant]['nom']
self.lig_donnees = db.data[self.item_courant]['nom']
def __init__(self, parent):
self._init_ctrls(parent)
#self._init_utils(parent)
#self.lig_donnees = ()
db = getdata.query("select nom,nom_commun,male,femelle from Popu")
index = 0
i=0
self.listCtrl1.Show(True)
nop=0
taillefish=(195,120,40,40,40)
for item in db.fields:
nop=nop+1
self.listCtrl1.InsertColumn(index,item[0].replace("_"," "),width=taillefish[(nop-1)])
index += 1
self.listCtrl1.InsertColumn((index+1), "Total",width=taillefish[(nop)])
nop=0
nopold=9
for row in range(len(db.data)):
nop=nop+1
for col in range(len(db.data[row])):
values = db.data[row][col]
if nop!=nopold :
nopold=nop
self.listCtrl1.InsertStringItem((nop-1), str(values))
else:
self.listCtrl1.SetStringItem(row, col, str(values))
self.listCtrl1.SetStringItem(row, (col+1), str(values+db.data[row][(col-1)]))
def _init_ctrls(self, prnt):
# generated method, don't edit
wx.Frame.__init__(self, id=wxID_MAINFRAME, name=u'Mainframe',
parent=prnt, pos=wx.Point(178, 126), size=wx.Size(794, 599),
style=wx.DEFAULT_FRAME_STYLE, title='Frame1')
self._init_utils()
self.SetClientSize(wx.Size(786, 571))
self.staticBitmap1 = wx.StaticBitmap(bitmap=wx.NullBitmap,
id=wxID_MAINFRAMESTATICBITMAP1, name='staticBitmap1', parent=self,
pos=wx.Point(16, 56), size=wx.Size(440, 272), style=0)
self.textCtrl1 = wx.TextCtrl(id=wxID_MAINFRAMETEXTCTRL1,
name='textCtrl1', parent=self, pos=wx.Point(488, 48),
size=wx.Size(100, 21), style=0, value='textCtrl1')
self.textCtrl2 = wx.TextCtrl(id=wxID_MAINFRAMETEXTCTRL2,
name='textCtrl2', parent=self, pos=wx.Point(680, 48),
size=wx.Size(100, 21), style=0, value='textCtrl2')
self.textCtrl3 = wx.TextCtrl(id=wxID_MAINFRAMETEXTCTRL3,
name='textCtrl3', parent=self, pos=wx.Point(488, 88),
size=wx.Size(100, 21), style=0, value='textCtrl3')
self.textCtrl4 = wx.TextCtrl(id=wxID_MAINFRAMETEXTCTRL4,
name='textCtrl4', parent=self, pos=wx.Point(680, 88),
size=wx.Size(100, 21), style=0, value='textCtrl4')
self.textCtrl5 = wx.TextCtrl(id=wxID_MAINFRAMETEXTCTRL5,
name='textCtrl5', parent=self, pos=wx.Point(680, 128),
size=wx.Size(100, 21), style=0, value='textCtrl5')
self.bitmapButton1 = wx.BitmapButton(bitmap=wx.NullBitmap,
id=wxID_MAINFRAMEBITMAPBUTTON1, name='bitmapButton1', parent=self,
pos=wx.Point(488, 128), size=wx.Size(24, 24),
style=wx.BU_AUTODRAW)
self.listCtrl1 = wx.ListCtrl(id=wxID_MAINFRAMELISTCTRL1,
name='listCtrl1', parent=self, pos=wx.Point(16, 336),
size=wx.Size(440, 208), style=wx.LC_ICON)
db = getdata.query("select * from Popu")
index = 0
i=0
#for i in db:
# self.listCtrl1.InsertStringItem(index, item[0])
# index += 1
self.listCtrl1.InsertStringItem(0, "George")
self.listCtrl1.SetStringItem(0, 1, "1234ABC")
self.listCtrl1.SetStringItem(0, 2, "100 Maple Ave.")
self.listCtrl1.SetStringItem(0, 3, "New York")
def OnListCtrl1LeftDclick(self, event):
#nomfiche = globalfunc.Fiche(self,"poisson",self.lig_donnees)
donn = self.lig_donnees
requete = "SELECT nom , descripteur , famille , synonyme , taille , origine , temperature , PH , durete , zone , description , vie , comportement , reproduction , dimorphisme , url FROM Fiche_POISSON WHERE nom LIKE '" + donn + "'"
db = getdata.query(requete)
donnees = db.data
msg="\n"
nop=0
for item in db.fields:
msg=msg + item[0].replace("_"," ") + " : "
try:
msg=msg + donnees[0][nop]
except :
msg=msg + "\n"
nop=nop+1
msg=msg + "\n\n"
dlg = wx.lib.dialogs.ScrolledMessageDialog(self, msg, "Fiche du Poisson")
dlg.Show()
from sklearn.feature_extraction.text import CountVectorizer
from sklearn.feature_extraction.text import TfidfTransformer
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn import svm
from sklearn import tree
from getdata import query
from sklearn.linear_model import LogisticRegression
from sklearn.ensemble import VotingClassifier
from sklearn.naive_bayes import MultinomialNB
from sklearn.ensemble import RandomForestClassifier
from sklearn.neighbors import KNeighborsClassifier
if __name__ == '__main__':
sql1 = "select * from trainmessage limit 0,700" #训练数据
sql2 = "select * from trainmessage limit 699,300" #测试数据
train = query(sql1)
test = query(sql2)
count_v1 = CountVectorizer()
counts_train = count_v1.fit_transform(train[0]) # 得到词频矩阵
transformer = TfidfTransformer(
) # Transform a count matrix to a normalized tf or tf-idf representation
tfidf_train = transformer.fit_transform(counts_train) # 得到频率矩阵
counts_test = count_v1.transform(test[0]) # fit_transform是将文本转为词频矩阵
# print(counts_test)
tfidf_test = transformer.transform(counts_test) # fit_transform是计算tf-idf
####KNN算法
knnclf = KNeighborsClassifier()
knnclf.fit(tfidf_train, train[1])
# -*- coding: cp1252 -*-
import re, urllib, getdata, html2text, time
#time.sleep(60)
rown=0
casereq=("descripteur","famille","synonyme","taille","origine","temperature","PH","durete","zone","description","vie","comportement","reproduction","dimorphisme")
while rown < 1199 :
#while rown < 938 :
rown=rown+1
requete = getdata.query("select url from Fiche_POISSON where id = " + str(rown))
bcl=0
htmlsanschar="<strong>Désolé 82.xxx.xxx.xxx, le site est en cours de maintenance...</strong><br/><small>Revenez dans 30 secondes.</small>"
bibi = re.findall('</strong><br/><small>(.*?)</small>',htmlsanschar)
try:
while bibi[0] == "Revenez dans 30 secondes." :
print requete.data[0][0]
htmlSource = urllib.urlopen(requete.data[0][0]).read()
htmlsanschar = htmlSource.replace("\n"," ")
print htmlsanschar
bibi = re.findall('</strong><br/><small>(.*?)</small>',htmlsanschar)
try:
if bibi[0] == "Revenez dans 30 secondes." :
print "sleeping begin at id " + str(rown)
time.sleep(31)
print "sleeping end"
except :
print "IOerror"
htmlSource = urllib.urlopen("http://www.aquabase.org/fish/dump.php3").read()
listeDeLiens = re.findall('">(.*?)</A>', htmlSource)
listeDeLiensurl = re.findall('<A HREF="(.*?)">', htmlSource)
nop = 0
for lien in listeDeLiens:
nop = nop + 1
if nop != 1:
# print lien
# print listeDeLiensurl[(nop-1)]
print(
"INSERT INTO Fiche_POISSON ( id , nom , url ) VALUES ( "
+ str(nop - 1)
+ " , '"
+ lien
+ "' , 'http://www.aquabase.org"
+ listeDeLiensurl[(nop - 1)]
+ "' ) "
)
getdata.query(
"INSERT INTO Fiche_POISSON ( id , nom , url ) VALUES ( "
+ str(nop - 1)
+ " , '"
+ lien
+ "' , 'http://www.aquabase.org"
+ listeDeLiensurl[(nop - 1)]
+ "' ) "
)
