reformatting dates with square brackets - regex

Question

When transferring my XML-files to TeX I try to reformat dates – my publisher said I had to use a smaller horizontal space between dates – and can't manage to go beyond the first step.
My input file is this
<a>
<date>January 1900</date>
<date>2. 2. 1902</date>
<date>3. [3]. 1903</date>
<date>[4. 4. 1904]</date>
</a>
where brackets mean that the date is not certain. There are all possible combination of brackets, e.g. second number of the year: 1[9]00. I created a command \mini which makes small space in TeX:
\newcommand{\mini}{\,}
The result after the xslt should be:
January 1900
2.{\mini}2.{\mini}1902
3.{\mini}[3].{\mini}1903
[4.{\mini}4.{\mini}1904]
I wrote a function, which tries to extract the square brackets and store their position to a variable and afterwards concat them back again. But as I don't manage to get the variables to show correct positions, I'm stuck:
<xsl:function name="foo:date-translate">
<xsl:param name="date-string" as="xs:string"/>
<xsl:variable name="opening-square-bracket" as="xs:integer" select="count(substring-before($date-string,'['))"/>
<xsl:variable name="closing-square-bracket" as="xs:integer" select="count(substring-before($date-string,'['))"/>
<xsl:variable name="date-string-without-square-brackets" as="xs:string" select="replace(replace($date-string,'\[',''),'\]','')"/>
<xsl:choose>
<xsl:when test="matches($date-string-without-square-brackets,'\d{1,2}. \d{1,2}. \d{4}')">
<xsl:choose>
<xsl:when test="not(contains($date-string,'['))">
<xsl:value-of select="replace($date-string,'(\d{1,2}). (\d{1,2}). (\d{4})','$1\\mini$2\\mini$3')"/>
</xsl:when>
<xsl:otherwise>
<xsl:value-of select="concat(substring(replace($date-string-without-square-brackets,'(\d{1,2}). (\d{1,2}). (\d{4})','$1\\mini$2\\mini$3'),0,$opening-square-bracket),'[',substring(replace($date-string-without-square-brackets,'(\d{1,2}). (\d{1,2}). (\d{4})','$1\\mini$2\\mini$3'),$opening-square-bracket, $closing-square-bracket))"/>
</xsl:otherwise>
</xsl:choose>
</xsl:when>
<xsl:otherwise>
<xsl:value-of select="$date-string"/>
</xsl:otherwise>
</xsl:choose>
</xsl:function>

Answer 1

I finally managed to solve it with recursion. Basically I have the test whether it suits the regex dd. mm. yyyy when the brackets are removed. as this works i can now rebuild the whole string.
<xsl:function name="foo:date-repeat">
<xsl:param name="date-string" as="xs:string"/>
<xsl:param name="amount" as="xs:integer"/>
<xsl:param name="counter" as="xs:integer"/>
<xsl:choose>
<xsl:when test="substring($date-string,$counter,1) =' '">
<xsl:text>\mini</xsl:text>
</xsl:when>
<xsl:otherwise>
<xsl:value-of select="substring($date-string,$counter,1)"/>
</xsl:otherwise>
</xsl:choose>
<xsl:if test="$counter <= $amount">
<xsl:value-of select="foo:date-repeat($date-string, $amount,$counter+1)"/>
</xsl:if>
</xsl:function>
<xsl:function name="foo:date-translate">
<xsl:param name="date-string" as="xs:string"/>
<xsl:variable name="date-string-without-square-brackets" as="xs:string" select="replace(replace($date-string,'\[',''),'\]','')"/>
<xsl:choose>
<xsl:when test="matches($date-string-without-square-brackets,'\d{1,2}. \d{1,2}. \d{4}')">
<xsl:choose>
<xsl:when test="not(contains($date-string,'['))"> <!-- Daten ohne eckige Klammer -->
<xsl:value-of select="replace($date-string,'(\d{1,2}). (\d{1,2}). (\d{4})','$1\\mini$2\\mini$3')"/>
</xsl:when>
<xsl:otherwise>
<xsl:value-of select="foo:date-repeat($date-string, string-length($date-string),1)"/>
</xsl:otherwise>
</xsl:choose>
</xsl:when>
<xsl:otherwise>
<xsl:value-of select="$date-string"/>
</xsl:otherwise>
</xsl:choose>
</xsl:function>
I have even expanded my solution, now the function checks tests every space ' ' if before is a dot or a dot and a bracket and afterwards is a number or a bracket and a number:
<xsl:function name="foo:date-repeat">
<xsl:param name="date-string" as="xs:string"/>
<xsl:param name="amount" as="xs:integer"/>
<xsl:param name="counter" as="xs:integer"/>
<xsl:choose>
<xsl:when test="substring($date-string,$counter,1) =' ' and ((substring($date-string,$counter -1,1) = '.' and number(substring($date-string,$counter -2,1)) = number(substring($date-string,$counter -2,1))) or (substring($date-string,$counter -2,2) = '.]' and number(substring($date-string,$counter -3,1)) = number(substring($date-string,$counter -3,1))))">
<xsl:choose>
<xsl:when test="number(substring($date-string,$counter +1,1)) = number(substring($date-string,$counter +1,1))">
<xsl:text>\mini</xsl:text>
</xsl:when>
<xsl:when test="substring($date-string,$counter +1,1) ='[' and number(substring($date-string,$counter +2,1)) = number(substring($date-string,$counter +2,1))">
<xsl:text>\mini</xsl:text>
</xsl:when>
<xsl:otherwise>
<xsl:value-of select="substring($date-string,$counter,1)"/>
</xsl:otherwise>
</xsl:choose>
</xsl:when>
<xsl:when test="substring($date-string,$counter,1) ='['">
<xsl:text>{[}</xsl:text>
</xsl:when>
<xsl:when test="substring($date-string,$counter,1) =']'">
<xsl:text>{]}</xsl:text>
</xsl:when>
<xsl:otherwise>
<xsl:value-of select="substring($date-string,$counter,1)"/>
</xsl:otherwise>
</xsl:choose>
<xsl:if test="$counter <= $amount">
<xsl:value-of select="foo:date-repeat($date-string, $amount,$counter+1)"/>
</xsl:if>