changeset 54:191c95187e87

working now on enron1k, uli1/sheet1
author Henry S. Thompson <ht@markup.co.uk>
date Tue, 16 May 2017 19:54:19 +0100
parents 8000dbed09df
children 05cf88c20cc5
files notes.txt tokenise.xsl
diffstat 2 files changed, 11 insertions(+), 6 deletions(-) [+]
line wrap: on
line diff
--- a/notes.txt	Tue May 16 19:09:03 2017 +0100
+++ b/notes.txt	Tue May 16 19:54:19 2017 +0100
@@ -225,13 +225,18 @@
 @@ string identify, to say nothing of actual value, is lost -- fix?
 @@ row/column/both spans
 ------
-enron/kenneth_lay__19506 contains this formula:
+enron1k/kenneth_lay__19506 contains this formula:
 
   <f>[1]!'.SPX'</f>
 
-which crashes tokenise/rnf
+which crashes tokenise/rnf FIXED works now, and with
+<f>[1]!'AES,DIVIDEND' (where _are_ these coming from???)
+Als, with enlarged memory, now runs on uli1/sheet1
 
 Changes intended to fix this fixed a bug (?) which wasn't properly
 merging e.g. +3 -- no examples of larger numbers available to check
-with...
+with...  We are now getting e.g. <x>2.509+0.482+0.238</x>
+in enron1k/kenneth_lay__19506
 
+We could _either_ add a class of operators, or a class of numbers?
+
--- a/tokenise.xsl	Tue May 16 19:09:03 2017 +0100
+++ b/tokenise.xsl	Tue May 16 19:54:19 2017 +0100
@@ -114,12 +114,12 @@
   <xsl:sequence select="[$content?1,$x/*]"/>
  </xsl:function>
 
- <xsl:function name="e:oddx">
+ <xsl:function name="e:oddx">  
   <xsl:param name="tokens" as="element(xf:group)*"/>
   <xsl:param name="i" as="xs:integer"/>
   <xsl:param name="soFar" as="xs:string"/><!-- Will always be a lonely ' -->
-  <!-- Amalgamate bits of type 9 and 10 as <v> for wrapping in <e> -->
-  <xsl:variable name="content" select="e:amalgamate($tokens,$i,$soFar,(10,9),$soFar)"/>
+  <!-- Amalgamate bits of type 3, 9 and 10 as <v> for wrapping in <e> -->
+  <xsl:variable name="content" select="e:amalgamate($tokens,$i,$soFar,(10,9,3),$soFar)"/>
   <xsl:variable name="x"><v><xsl:value-of select="
     if ($tokens[$content?1]=$soFar)
     then concat($content?2,$soFar)