From a80630656d5b8fb8fc640908259ab38d0ef9a34c Mon Sep 17 00:00:00 2001
From: Kevin Brubeck Unhammer <pixiemotion@gmail.com>
Date: Mon, 9 Jun 2008 14:09:24 +0200
Subject: [PATCH] todo: test PSTOP(h|ln)

---
 src/dmv.py  |  38 +++++++++-----------------------------
 src/dmv.pyc | Bin 16490 -> 16448 bytes
 src/io.pyc  | Bin 6735 -> 6846 bytes
 3 files changed, 9 insertions(+), 29 deletions(-)

diff --git a/src/dmv.py b/src/dmv.py
index 05ee6e8..ae39a98 100755
--- a/src/dmv.py
+++ b/src/dmv.py
@@ -404,14 +404,17 @@ def reestimate_zeros(h_nums):
     return f
 
 def reestimate(g, corpus):
-    '''P_STOP(-STOP|...) = 1 - P_STOP(STOP|...) '''
+    '''current todo.
+    P_STOP(-STOP|...) = 1 - P_STOP(STOP|...) '''
     f = reestimate_zeros(g.head_nums)
     for sent in corpus:
         chart = {}
         sent_nums = [g.tagnum(w) for i,w in enumerate(sent)]
         for loc_h,h in zip(range(len(sent)), sent_nums):
             inner_sent(loc_h, g, sent,      chart)
+            io.debug( debug_chart (g,sent,chart) ,'reest_chart')
             prune_sent(loc_h, g, sent_nums, chart)
+            io.debug( debug_chart( g,sent,chart) ,'reest_chart')
             for s in range(loc_h): # s<loc(h), range gives strictly less
                 for t in range(loc_h, len(sent)):
                     io.debug( "s:%s t:%s loc:%d"%(s,t,loc_h)  , 'reest')
@@ -421,13 +424,14 @@ def reestimate(g, corpus):
                     if (s, t, (RBAR,h), loc_h) in chart:
                         f[('STOP','den',h)] += chart[(s, t, (RBAR,h), loc_h)]
                         io.debug( "den+=%s"%chart[(s, t, (RBAR,h), loc_h)]  , 'reest')
+                    io.debug( "num:%s den:%s"%(f[('STOP','num',h)],f[('STOP','den',h)]), 'reest')
             for s in [loc_h]:
                 # adjacent left attachment
                 pass
         
         # todo: use sum([chart[(s, t...)] etc? but can we then
         # keep den and num separate within _one_ sum()-call? use map?
-
+            
     for r in g.all_rules():
         if r.L() == STOP:
             if f[('STOP','den',h)] > 0.0:
@@ -460,24 +464,7 @@ def testreestimation():
                                       'det nn vbd',      'det nn vbd c vbd pp', 
                                       'det nn vbd pp',   'det nn vbd det', ]]
     g = harmonic.initialize(corpus)
-    
-    print "takes some time"
-    for r in g.h_rules(h):
-        if r.L()==STOP:
-            print r
-#             print "off-set the rule, see what happens:"
-#             r.probN = 0.7
-#             print r
-            for i in range(3):
-                pstophln = reestimate(g, corpus)
-                print "p(STOP|%s,L,N):%s"%(h_tag,pstophln)
-                
-                for r in g.h_rules(h):
-                    if r.L()==STOP:
-                        print r
-                        r.probN = pstophln
-                        print r
-    return None
+    reestimate(g, corpus)
 
 def testgrammar_h():                            # Non, Adj
     _h_ = DMV_Rule((LRBAR,0), STOP,    ( RBAR,0), 1.0, 1.0) # LSTOP
@@ -504,16 +491,9 @@ def testprune():
     print debug_chart(g,sent,chart)
 
 def testreestimation_h():
+    io.DEBUG=['reest']
     g = testgrammar_h()
-    for r in g.h_rules(0):
-        if r.L()==STOP:
-            print r
-            for i in range(9):
-                reestimate(g, ['h h h'.split()])
-                for r in g.h_rules(0):
-                    if r.L()==STOP:
-                        print r
-    print "doing something obviously wrong here..."
+    reestimate(g,['h h h'.split()])
 
 if __name__ == "__main__":
     io.DEBUG = []
diff --git a/src/dmv.pyc b/src/dmv.pyc
index 5c0b077101fcbf2325aff1c92b953decd88caa53..0864141ee5e1e752e89b4c986dbda98a4b24c09b 100644
GIT binary patch
delta 1170
zcwU85OHUI~6h3EWI%P^*il!t8NH>ORB99h?K(R!PiHX5j7mW`@!nCwRJG8x+E@o_#
zCQ97io&E`xKS1I_<63_J3ti~KL|wT+&zb3fCMccWFL%!Q&UYSn?&lWnv{2vv773+?
zUh1#Lhc;9I0bm1^5!wij0*#VlBV1a#r^pZE3(WyVXdnV0bg)&X(BJNYwnCIWyrY0)
zfD-}{1lwR8BH0?WH9*k?WdlU$5333$awv*!pWgL<dd}lF9Hbg3q|K#S=yXG)*fAYc
z#x-3-Ck&pJseny+!b~cv&}XWymRwH7Wi}PnG%y45dMX+8*t%?`1}8jD1p&QS5dk{{
zMa31-xU6*@tR-?QV3I<R6ynXlN&VAmrc;93-l0nUem$J4!={mVssWd(LcgEIf2VLi
zt{TLA^$X-1z0f4dHJ*$PG?{8i$yWvBt9|7B>K<+s+Cc7IaHxm~lzaGH#9&<k>+L}=
zeikzX>l#??ea}DZiIduLZ_}T6H`|YjIm#cW>FhUIQvKnC57YMMan|ELQF!Ie2;4|U
z0iZrdu)0~R6)GYjEYlhdCFm#LrBBD`3k0YY3U#r(T3V?I`8|C(R*#UeviV^4R6Su9
zDzkEWauJWo$CG2R6XeN_Gu&n<$eqdO=?k<4wFiTsqmJir5W;vI2i+ARTOX|<9L6ym
z?O<kmYwDMRmfV^d!HRq_a|PdQ|C}i*_(`tJz1M9P-;i^eMf@S3W(Jzb8E9SPe#Qvs
z2GLQ|rClae_#7QNN|le9KKQc6mNoCzm_UQz;Kkg14i6BN67(c_mvdXX#o%VO+|&d~
z&Y8Q(JCGme&qVjgP}jLv45J|@&hN&nOqpWHFw8SNlq1=@aaU$@U)5oFB%fuIxFKI>
STgHSjY($Na5s+6GPW}e*bk|A%

delta 1237
zcwU85OKTHR6h7xBnPieo+gK7vAG8#-X`yNQlw$e-i)}%>h>K8sft|FS(sq(&su5>M
zB=`rsAc$*ME{gUK2yWcD(S^$@xNxB&uC$`(Op*#wk(v8(?|I(yopYa6@IwWagD+Yz
z8+)QWO?7)c02Y7^U?5Zx908>xv=Qd7d`0;>HFtd<EiYI;uoSR8Y(l>sfT|#qdw5xZ
zqk<Cz%MZ57amdK_Le&ds1sbA2LsVKq3PsW?XE=jS2uh=Je<V8CiB1?S6^tP0lpq9l
znC;d)5Yq)`!<x(r$&VwQD<06jU<X;hDj-hb8){c7TwgE5PkYk!pg#DgQ6s|+0?N73
zsQf<CLd58b)J7Bjn!8`7N4v^3yTVI%WeW$8Ek!uWI+vpWl~XvvY{4R3jC7w|Vl}7&
zsSm-?z}CQM;j`5Wn*z)iRF?pqlhw<K>jBvGf>}M*GzyK|<tEc8g}c_)piO1*A3jYk
z|I<f_kOm)xBZ2kUK{s})$(blTw(S(WvU&(ECc*$v93YU3;{)=}_&NDz{5(eFsfpoG
zj>MS`hLp@rJRC~X7DOxhK|uv4u?xc3PM=2S7_DyX!C_3gIC(HO`BUJI9GmLLZFy;G
z37;Q4npzY1R)*8Bl-nHEl<(8a<<Cf0XyMqXVi*<Cg%Lc7t1O|J2Oq7oe9gz;z5r^8
z15mnOe+oL*M)WQ?G>r&`aGU0Fo3;h~LKYTFI?&r-0@yUekB6rlh@>(Y3VYCCFgvBO
zrnr57i>ub-7V^Yro@QvCR-E>6$P(!bi#iaweDh_4>o8O<fX^Gf^cDMQr<=ykdT(*X
z+|F4Wyfv@iFwDK;)~()~yC$7Ca=W=?GMVr*S1w`9Qsx9hgy90ij>{xLg2~FPFoYNu
zWF;Dx@1|pY-<aW!I>T+$sdn_$x>i5MrGMrynrE*G2JUobnqg1=n7JBtb6CgMuo)bg
ipG}}CAIw(zhtwXmMGdN9&21qi5vovekQ%9p-}nnU3*Wo|

diff --git a/src/io.pyc b/src/io.pyc
index 10b2c64a8853e9c931829e61eb43a33c3f276ad5..920f57e60d4c2485ec8aa752725b73785f159e8b 100644
GIT binary patch
delta 794
zcwS{vO=}ZT6rGpJOeUF+*oJ(})OJd1&DdyT8WXJxD{7%dO66G)6$Hf^W?@pBDs&?h
z1rY=ncV(!!aIH|rjT^zeB7%QH=t8n^<HmPpm;~_+9NxX}ymQaJ^S+(>)JO$=GxOS5
z-A{d&*(I7tFD3a^FBRCyN`i`3nGVz}Jyuin#v0FTCnZU0KfK?q(t_GlxFJG%s@B-N
zh?nS(dYUz8*{IXK)CYEm#<W>ICn!7%7^N%Pk=z-v?P>xfInD5SX?mtL^X8BT$P4f!
z`el@Is5yySmH<^#Sy~Q$YkkI6=x1gp?|?oAdICEYa9ETo(MvtUE>ll`pyHT2G-5Q_
zs_^&Rb-`B#d@a`Cp_+@j3M{OVX?B@QyJnTG)4o}!o;9t2evJ&P&Zg<SHKpL@brIiK
zxe1@}Au&&fQ~^)~lmO#^3BY9BC7=1<LDacoi|xgroV_cvPVi#%oxB`QQ~!s%jW)J!
zw_6*bQdIUR4j3R(M!!5WyhpnQht1Gl!L_}x9(+T{@62uP@G2zfN8!B#Azm3cJP9ZR
zYP458cA`2+)&O37Jk$#9`R(?`)-CE6J5%uJBiHJ^!uq(_`uYXyV39XM>-41bb`H-Z
z{Gq;#<S@i*h#F1AO1weqcI5)5#hU;OJ?#EYI0(J~IEy629&OG?mS<y{AC2@leYV%c
JFCB5t{ROKlhtU84

delta 745
zcwSYKO-mb56o%(I@y<+UGHIh`CZmnP+9nzrgNfSGQu+ZwYD?)=q|imM7$hKxRS`rN
zf=ej~2fFF9>wd^=rJMeM)?ZO5luEYU^uBi*gA6>JbMN_h&$;t;=6lr$JBheK3n#{z
z&hFA%>t0MO7I-)oEB4VTFVZ$o({Em;FMN)cj2fG!_r_0_q(gJkPHCDZT;M7lbDw^g
z&cKXxw@e95w-~;HNq@}QD^VCQC{fU4f|n98Y3GpPNN|(0uwiu=tI-J`q0_iWM|N!}
zgRB5qeXuw%tQsC0F{55@HR|<wvYa;@K_Wv3&MbQ(bynz?v&^#eFfm|9S3nl6xsvge
z_7ht&;5;$KUQo`hupE7M$NHgviT2zIS;;(GR@S0>{TcFKLN0`8tYD+w5Re8uU<9}Z
zjCQ9L<AB*6L&=CXz=bdE2jRQqs?OHKzp0P<yxOXLDR{eIVsoXn)=;IQzBh2efT(m9
zi1+g7JRPucvc1tvUNwYp6l%6h8@r+;tu{UOJ_RTwCNCsT0XKm%+5W_>(#2sJ$ak-M
zwUoSOtJT<iMJs+=Ha+u)c~yy|j{l)Fr~QAtrjEm&i}#O4QZDL<Dt*foA7L*7Ka<+8
kn)9_9S=<Kh;1I=jHl`zS6Wy?8w7epfgXi)+4}$yu0KDaYegFUf

-- 
2.11.4.GIT