From 6de3ca3891d4b1e35930749f76012d42765eed2c Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?=E2=80=9CNaeel=E2=80=9D?= <tazet@narod.ru>
Date: Thu, 18 Jun 2026 18:19:06 +0400
Subject: [PATCH] =?UTF-8?q?v1.0.12:=20parser.cfm=20=E2=80=94=20=D1=80?=
 =?UTF-8?q?=D0=B5=D1=84=D0=B0=D0=BA=D1=82=D0=BE=D1=80=D0=B8=D0=BD=D0=B3=20?=
 =?UTF-8?q?+=20ZIP=20(=D0=BF=D0=BE=D0=B4=D0=BF=D0=B0=D0=BF=D0=BA=D0=B8,=20?=
 =?UTF-8?q?PDF/DOCX)?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 index.cfm  |   2 +-
 parser.cfm | 411 +++++++++++++++++++++++++++++------------------------
 2 files changed, 229 insertions(+), 184 deletions(-)

diff --git a/index.cfm b/index.cfm
index 6e7d5a7..5e57deb 100644
--- a/index.cfm
+++ b/index.cfm
@@ -1 +1 @@
-<cfsetting showdebugoutput=no enablecfoutputonly=yes><cfoutput>OK v1.0.11</cfoutput>
+<cfsetting showdebugoutput=no enablecfoutputonly=yes><cfoutput>OK v1.0.12</cfoutput>
diff --git a/parser.cfm b/parser.cfm
index 345d830..f863738 100644
--- a/parser.cfm
+++ b/parser.cfm
@@ -3,202 +3,256 @@
 
 <cfparam name="url.doc_id" default="">
 
-<cfset result = {ok: false, error: ""}>
+<!--- ═══════════════════════════════════════════════════════════
+     Helper: parsePDF — PDFBox → {elements, errors}
+     ═══════════════════════════════════════════════════════════ --->
+<cffunction name="parsePDF" access="private" returntype="struct">
+    <cfargument name="fileBytes" type="binary">
+    <cfset var out = {elements: [], errors: []}>
+    <cftry>
+        <cfset var bais = createObject("java", "java.io.ByteArrayInputStream").init(fileBytes)>
+        <cfset var bis = createObject("java", "java.io.BufferedInputStream").init(bais)>
+        <cfset var pdfDoc = 0>
+        <cftry>
+            <cfset pdfDoc = createObject("java", "org.apache.pdfbox.Loader").loadPDF(bis)>
+            <cfcatch>
+                <cftry>
+                    <cfset pdfDoc = createObject("java", "org.apache.pdfbox.pdmodel.PDDocument").load(bis)>
+                    <cfcatch>
+                        <cfset arrayAppend(out.errors, "PDFBox: cannot open — " & cfcatch.message)>
+                    </cfcatch>
+                </cftry>
+            </cfcatch>
+        </cftry>
+        <cfif NOT isNumeric(pdfDoc)>
+            <cfset var stripper = createObject("java", "org.apache.pdfbox.text.PDFTextStripper")>
+            <cfset var pageCount = pdfDoc.getNumberOfPages()>
+            <cfloop from="1" to="#pageCount#" index="var pn">
+                <cfset stripper.setStartPage(pn)>
+                <cfset stripper.setEndPage(pn)>
+                <cfset var pageText = stripper.getText(pdfDoc)>
+                <cfif len(trim(pageText))>
+                    <cfloop list="#pageText#" index="var line" delimiters="#chr(10)#">
+                        <cfset line = trim(line)>
+                        <cfif len(line)>
+                            <cfset arrayAppend(out.elements, {type: "paragraph", style: "", text: line, page: pn})>
+                        </cfif>
+                    </cfloop>
+                </cfif>
+            </cfloop>
+            <cfset pdfDoc.close()>
+        </cfif>
+        <cfset bis.close()>
+        <cfset bais.close()>
+        <cfcatch>
+            <cfset arrayAppend(out.errors, "PDFBox: " & cfcatch.message)>
+        </cfcatch>
+    </cftry>
+    <cfreturn out>
+</cffunction>
 
+<!--- ═══════════════════════════════════════════════════════════
+     Helper: parseDOCX — ZIP/XML → {elements, errors}
+     ═══════════════════════════════════════════════════════════ --->
+<cffunction name="parseDOCX" access="private" returntype="struct">
+    <cfargument name="fileBytes" type="binary">
+    <cfset var out = {elements: [], errors: []}>
+    <cftry>
+        <cfset var bais = createObject("java", "java.io.ByteArrayInputStream").init(fileBytes)>
+        <cfset var zis = createObject("java", "java.util.zip.ZipInputStream").init(bais)>
+        <cfset var xmlStr = "">
+        <cfloop condition="true">
+            <cfset var entry = zis.getNextEntry()>
+            <cfif isNull(entry)><cfbreak></cfif>
+            <cfif entry.getName() EQ "word/document.xml">
+                <cftry>
+                    <cfset var xmlBytes = zis.readAllBytes()>
+                    <cfset xmlStr = createObject("java", "java.lang.String").init(xmlBytes, "UTF-8")>
+                    <cfcatch>
+                        <cfset var baos = createObject("java", "java.io.ByteArrayOutputStream").init()>
+                        <cfloop condition="true">
+                            <cfset var b = zis.read()>
+                            <cfif b LT 0><cfbreak></cfif>
+                            <cfset baos.write(b)>
+                        </cfloop>
+                        <cfset xmlStr = createObject("java", "java.lang.String").init(baos.toByteArray(), "UTF-8")>
+                    </cfcatch>
+                </cftry>
+                <cfbreak>
+            </cfif>
+            <cfset zis.closeEntry()>
+        </cfloop>
+        <cfset zis.close()>
+        <cfset bais.close()>
+        <cfif NOT len(xmlStr)>
+            <cfset arrayAppend(out.errors, "word/document.xml not found")>
+        <cfelse>
+            <cfset var docXml = xmlParse(xmlStr)>
+            <cfset var body = "">
+            <cftry>
+                <cfset body = xmlSearch(docXml, "//:body")>
+                <cfcatch>
+                    <cfset body = xmlSearch(docXml, "//*[local-name()='body']")>
+                </cfcatch>
+            </cftry>
+            <cfif arrayLen(body) GT 0>
+                <cfset body = body[1]>
+                <cfloop array="#body.XmlChildren#" index="var child">
+                    <cfset var localName = child.XmlName>
+                    <cfif localName EQ "p" OR right(localName, 2) EQ ":p">
+                        <cfset var paraText = "">
+                        <cftry>
+                            <cfset var textNodes = xmlSearch(child, "//:t")>
+                            <cfcatch>
+                                <cfset textNodes = xmlSearch(child, "//*[local-name()='t']")>
+                            </cfcatch>
+                        </cftry>
+                        <cfloop array="#textNodes#" index="var tn">
+                            <cfset paraText &= tn.XmlText>
+                        </cfloop>
+                        <cfset paraText = trim(paraText)>
+                        <cfif len(paraText)>
+                            <cfset arrayAppend(out.elements, {type: "paragraph", style: "", text: paraText, page: 0})>
+                        </cfif>
+                    <cfelseif localName EQ "tbl" OR right(localName, 4) EQ ":tbl">
+                        <cfset var rows = []>
+                        <cftry>
+                            <cfset var trNodes = xmlSearch(child, "//:tr")>
+                            <cfcatch>
+                                <cfset trNodes = xmlSearch(child, "//*[local-name()='tr']")>
+                            </cfcatch>
+                        </cftry>
+                        <cfloop array="#trNodes#" index="var trNode">
+                            <cfset var row = []>
+                            <cftry>
+                                <cfset var tcNodes = xmlSearch(trNode, ".//:tc")>
+                                <cfcatch>
+                                    <cfset tcNodes = xmlSearch(trNode, ".//*[local-name()='tc']")>
+                                </cfcatch>
+                            </cftry>
+                            <cfloop array="#tcNodes#" index="var tcNode">
+                                <cfset var cellText = "">
+                                <cftry>
+                                    <cfset var tNodes = xmlSearch(tcNode, ".//:t")>
+                                    <cfcatch>
+                                        <cfset tNodes = xmlSearch(tcNode, ".//*[local-name()='t']")>
+                                    </cfcatch>
+                                </cftry>
+                                <cfloop array="#tNodes#" index="var tNode">
+                                    <cfset cellText &= tNode.XmlText>
+                                </cfloop>
+                                <cfset arrayAppend(row, trim(cellText))>
+                            </cfloop>
+                            <cfif arrayLen(row) GT 0>
+                                <cfset arrayAppend(rows, row)>
+                            </cfif>
+                        </cfloop>
+                        <cfif arrayLen(rows) GT 0>
+                            <cfset arrayAppend(out.elements, {type: "table", rows: rows, page: 0})>
+                        </cfif>
+                    </cfif>
+                </cfloop>
+            <cfelse>
+                <cfset arrayAppend(out.errors, "body not found")>
+            </cfif>
+        </cfif>
+        <cfcatch>
+            <cfset arrayAppend(out.errors, "DOCX: " & cfcatch.message)>
+        </cfcatch>
+    </cftry>
+    <cfreturn out>
+</cffunction>
+
+<!--- ═══════════════════════════════════════════════ MAIN --->
+<cfset result = {ok: false, error: ""}>
 <cftry>
     <cfif NOT len(url.doc_id)>
         <cfset result.error = "doc_id required">
     <cfelse>
-        <!--- Читаем документ из БД --->
         <cfquery name="doc" datasource="baza">
             SELECT id, filename, mime_type, original_bytes, status
-            FROM documents
-            WHERE id = <cfqueryparam value="#url.doc_id#" cfsqltype="cf_sql_varchar">
+            FROM documents WHERE id = <cfqueryparam value="#url.doc_id#" cfsqltype="cf_sql_varchar">
         </cfquery>
-
         <cfif doc.recordCount EQ 0>
             <cfset result.error = "document not found">
         <cfelse>
             <cfset elements = []>
             <cfset errors = []>
+            <cfset files = []>
 
-            <!--- PDF --->
-            <cfif doc.mime_type EQ "application/pdf">
+            <!--- ZIP: поиск PDF/DOCX во всех подпапках --->
+            <cfif doc.mime_type EQ "application/zip">
                 <cftry>
-                    <!--- Base64 → байты --->
-                    <cfset fileBytes = binaryDecode(doc.original_bytes, "base64")>
-                    <!--- Байты → InputStream --->
-                    <cfset bais = createObject("java", "java.io.ByteArrayInputStream").init(fileBytes)>
-                    <cfset bis = createObject("java", "java.io.BufferedInputStream").init(bais)>
-
-                    <!--- Пробуем PDFBox 3.x Loader, иначе 2.x PDDocument.load --->
-                    <cfset pdfDoc = 0>
-                    <cftry>
-                        <cfset pdfDoc = createObject("java", "org.apache.pdfbox.Loader").loadPDF(bis)>
-                        <cfcatch>
-                            <cftry>
-                                <cfset pdfDoc = createObject("java", "org.apache.pdfbox.pdmodel.PDDocument").load(bis)>
-                                <cfcatch>
-                                    <cfset arrayAppend(errors, "PDFBox: cannot open PDF — " & cfcatch.message)>
-                                </cfcatch>
-                            </cftry>
-                        </cfcatch>
-                    </cftry>
-                    <cfif NOT isNumeric(pdfDoc)>
-                        <cfset stripper = createObject("java", "org.apache.pdfbox.text.PDFTextStripper")>
-
-                        <cfset pageCount = pdfDoc.getNumberOfPages()>
-                        <cfloop from="1" to="#pageCount#" index="pn">
-                            <cfset stripper.setStartPage(pn)>
-                            <cfset stripper.setEndPage(pn)>
-                            <cfset pageText = stripper.getText(pdfDoc)>
-                            <cfif len(trim(pageText))>
-                                <cfloop list="#pageText#" index="line" delimiters="#chr(10)#">
-                                    <cfset line = trim(line)>
-                                    <cfif len(line)>
-                                        <cfset arrayAppend(elements, {type: "paragraph", style: "", text: line, page: pn})>
-                                    </cfif>
-                                </cfloop>
-                            </cfif>
-                        </cfloop>
-
-                        <cfset pdfDoc.close()>
-                    </cfif>
-                    <cfset bis.close()>
-                    <cfset bais.close()>
-
-                    <cfcatch>
-                        <cfset arrayAppend(errors, "PDFBox: " & cfcatch.message)>
-                    </cfcatch>
-                </cftry>
-
-            <!--- DOCX — ZIP/XML (POI не нужен) --->
-            <cfelseif doc.mime_type EQ "application/vnd.openxmlformats-officedocument.wordprocessingml.document">
-                <cftry>
-                    <cfset fileBytes = binaryDecode(doc.original_bytes, "base64")>
-                    <cfset bais = createObject("java", "java.io.ByteArrayInputStream").init(fileBytes)>
-                    <cfset zis = createObject("java", "java.util.zip.ZipInputStream").init(bais)>
-
-                    <cfset xmlStr = "">
+                    <cfset var zipBytes = binaryDecode(doc.original_bytes, "base64")>
+                    <cfset var zbais = createObject("java", "java.io.ByteArrayInputStream").init(zipBytes)>
+                    <cfset var zzis = createObject("java", "java.util.zip.ZipInputStream").init(zbais)>
                     <cfloop condition="true">
-                        <cfset entry = zis.getNextEntry()>
-                        <cfif isNull(entry)><cfbreak></cfif>
-                        <cfif entry.getName() EQ "word/document.xml">
-                            <cftry>
-                                <cfset xmlBytes = zis.readAllBytes()>
-                                <cfset xmlStr = createObject("java", "java.lang.String").init(xmlBytes, "UTF-8")>
-                                <cfcatch>
-                                    <!--- fallback: читаем через буфер --->
-                                    <cfset baos = createObject("java", "java.io.ByteArrayOutputStream").init()>
-                                    <cfloop condition="true">
-                                        <cfset b = zis.read()>
-                                        <cfif b LT 0><cfbreak></cfif>
-                                        <cfset baos.write(b)>
-                                    </cfloop>
-                                    <cfset xmlStr = createObject("java", "java.lang.String").init(baos.toByteArray(), "UTF-8")>
-                                </cfcatch>
-                            </cftry>
-                            <cfbreak>
+                        <cfset var zentry = zzis.getNextEntry()>
+                        <cfif isNull(zentry)><cfbreak></cfif>
+                        <cfset var zname = zentry.getName()>
+                        <cfif right(zname, 1) EQ "/">
+                            <cfset zzis.closeEntry()><cfcontinue>
                         </cfif>
-                        <cfset zis.closeEntry()>
-                    </cfloop>
-                    <cfset zis.close()>
-                    <cfset bais.close()>
-
-                    <cfif NOT len(xmlStr)>
-                        <cfset arrayAppend(errors, "DOCX: word/document.xml not found in ZIP")>
-                    <cfelse>
-                        <cfset docXml = xmlParse(xmlStr)>
-                        <cfset ns = "http://schemas.openxmlformats.org/wordprocessingml/2006/main">
-
-                        <!--- Ищем body: или с namespace или без --->
-                        <cfset body = "">
+                        <cfset var entryBytes = "">
                         <cftry>
-                            <cfset body = xmlSearch(docXml, "//:body")>
+                            <cfset entryBytes = zzis.readAllBytes()>
                             <cfcatch>
-                                <cfset body = xmlSearch(docXml, "//*[local-name()='body']")>
+                                <cfset var zbaos = createObject("java", "java.io.ByteArrayOutputStream").init()>
+                                <cfloop condition="true">
+                                    <cfset var zb = zzis.read()>
+                                    <cfif zb LT 0><cfbreak></cfif>
+                                    <cfset zbaos.write(zb)>
+                                </cfloop>
+                                <cfset entryBytes = zbaos.toByteArray()>
                             </cfcatch>
                         </cftry>
-
-                        <cfif arrayLen(body) GT 0>
-                            <cfset body = body[1]>
-                            <cfloop array="#body.XmlChildren#" index="child">
-                                <cfset localName = child.XmlName>
-                                <cfif localName EQ "p" OR right(localName, 2) EQ ":p">
-                                    <!--- Параграф: собираем текст из всех w:t --->
-                                    <cfset paraText = "">
-                                    <cftry>
-                                        <cfset textNodes = xmlSearch(child, "//:t")>
-                                        <cfcatch>
-                                            <cfset textNodes = xmlSearch(child, "//*[local-name()='t']")>
-                                        </cfcatch>
-                                    </cftry>
-                                    <cfloop array="#textNodes#" index="tn">
-                                        <cfset paraText &= tn.XmlText>
-                                    </cfloop>
-                                    <cfset paraText = trim(paraText)>
-                                    <cfif len(paraText)>
-                                        <cfset arrayAppend(elements, {type: "paragraph", style: "", text: paraText, page: 0})>
-                                    </cfif>
-                                <cfelseif localName EQ "tbl" OR right(localName, 4) EQ ":tbl">
-                                    <!--- Таблица --->
-                                    <cfset rows = []>
-                                    <cftry>
-                                        <cfset trNodes = xmlSearch(child, "//:tr")>
-                                        <cfcatch>
-                                            <cfset trNodes = xmlSearch(child, "//*[local-name()='tr']")>
-                                        </cfcatch>
-                                    </cftry>
-                                    <cfloop array="#trNodes#" index="trNode">
-                                        <cfset row = []>
-                                        <cftry>
-                                            <cfset tcNodes = xmlSearch(trNode, ".//:tc")>
-                                            <cfcatch>
-                                                <cfset tcNodes = xmlSearch(trNode, ".//*[local-name()='tc']")>
-                                            </cfcatch>
-                                        </cftry>
-                                        <cfloop array="#tcNodes#" index="tcNode">
-                                            <cfset cellText = "">
-                                            <cftry>
-                                                <cfset tNodes = xmlSearch(tcNode, ".//:t")>
-                                                <cfcatch>
-                                                    <cfset tNodes = xmlSearch(tcNode, ".//*[local-name()='t']")>
-                                                </cfcatch>
-                                            </cftry>
-                                            <cfloop array="#tNodes#" index="tNode">
-                                                <cfset cellText &= tNode.XmlText>
-                                            </cfloop>
-                                            <cfset arrayAppend(row, trim(cellText))>
-                                        </cfloop>
-                                        <cfif arrayLen(row) GT 0>
-                                            <cfset arrayAppend(rows, row)>
-                                        </cfif>
-                                    </cfloop>
-                                    <cfif arrayLen(rows) GT 0>
-                                        <cfset arrayAppend(elements, {type: "table", rows: rows, page: 0})>
-                                    </cfif>
-                                </cfif>
-                            </cfloop>
+                        <cfset var zext = listLast(zname, ".")>
+                        <cfset var parsed = {elements: [], errors: []}>
+                        <cfif zext EQ "pdf">
+                            <cfset parsed = parsePDF(entryBytes)>
+                        <cfelseif zext EQ "docx">
+                            <cfset parsed = parseDOCX(entryBytes)>
                         <cfelse>
-                            <cfset arrayAppend(errors, "DOCX: body not found in document.xml")>
+                            <cfset arrayAppend(parsed.errors, "unsupported: .#zext#")>
                         </cfif>
-                    </cfif>
-
+                        <cfset var prefix = doc.filename & "/">
+                        <cfloop array="#parsed.elements#" index="var el">
+                            <cfset el.archive_path = prefix & zname>
+                            <cfset arrayAppend(elements, el)>
+                        </cfloop>
+                        <cfset arrayAppend(files, {
+                            filename: zname, archive_path: prefix & zname,
+                            element_count: arrayLen(parsed.elements),
+                            error_count: arrayLen(parsed.errors)
+                        })>
+                        <cfloop array="#parsed.errors#" index="var e">
+                            <cfset arrayAppend(errors, zname & ": " & e)>
+                        </cfloop>
+                        <cfset zzis.closeEntry()>
+                    </cfloop>
+                    <cfset zzis.close()><cfset zbais.close()>
                     <cfcatch>
-                        <cfset arrayAppend(errors, "DOCX: " & cfcatch.message)>
+                        <cfset arrayAppend(errors, "ZIP: " & cfcatch.message)>
                     </cfcatch>
                 </cftry>
 
-            <!--- .doc — старый формат, пропускаем --->
-            <cfelseif doc.mime_type EQ "application/msword">
-                <cfset arrayAppend(errors, ".doc parsing not available (use libreoffice to convert)")>
+            <cfelseif doc.mime_type EQ "application/pdf">
+                <cfset var fileBytes = binaryDecode(doc.original_bytes, "base64")>
+                <cfset var parsed = parsePDF(fileBytes)>
+                <cfset elements = parsed.elements>
+                <cfset errors = parsed.errors>
+
+            <cfelseif doc.mime_type EQ "application/vnd.openxmlformats-officedocument.wordprocessingml.document">
+                <cfset var fileBytes = binaryDecode(doc.original_bytes, "base64")>
+                <cfset var parsed = parseDOCX(fileBytes)>
+                <cfset elements = parsed.elements>
+                <cfset errors = parsed.errors>
 
-            <!--- TXT — plain text --->
             <cfelseif doc.mime_type EQ "text/plain" OR doc.mime_type EQ "application/octet-stream">
                 <cftry>
-                    <cfset fileBytes = binaryDecode(doc.original_bytes, "base64")>
-                    <cfset text = createObject("java", "java.lang.String").init(fileBytes, "UTF-8")>
-                    <cfloop list="#text#" index="line" delimiters="#chr(10)#">
+                    <cfset var fileBytes = binaryDecode(doc.original_bytes, "base64")>
+                    <cfset var text = createObject("java", "java.lang.String").init(fileBytes, "UTF-8")>
+                    <cfloop list="#text#" index="var line" delimiters="#chr(10)#">
                         <cfset line = trim(line)>
                         <cfif len(line)>
                             <cfset arrayAppend(elements, {type: "paragraph", style: "", text: line, page: 1})>
@@ -209,16 +263,16 @@
                     </cfcatch>
                 </cftry>
 
+            <cfelseif doc.mime_type EQ "application/msword">
+                <cfset arrayAppend(errors, ".doc not supported (use libreoffice)")>
             <cfelse>
-                <cfset arrayAppend(errors, "unsupported mime_type: " & doc.mime_type)>
+                <cfset arrayAppend(errors, "unsupported: " & doc.mime_type)>
             </cfif>
 
-            <!--- Сохраняем результат --->
-            <cfset newStatus = arrayLen(elements) GT 0 ? "parsed" : "error">
+            <cfset var newStatus = arrayLen(elements) GT 0 ? "parsed" : "error">
             <cfif arrayLen(errors) GT 0 AND arrayLen(elements) EQ 0>
                 <cfset newStatus = "error">
             </cfif>
-
             <cfquery datasource="baza">
                 UPDATE documents
                 SET elements_json = <cfqueryparam value="#serializeJSON(elements)#" cfsqltype="cf_sql_varchar">::jsonb,
@@ -226,23 +280,14 @@
                     error_message = <cfqueryparam value="#arrayLen(errors) ? serializeJSON(errors) : ''#" cfsqltype="cf_sql_varchar">
                 WHERE id = <cfqueryparam value="#url.doc_id#" cfsqltype="cf_sql_varchar">
             </cfquery>
-
-            <cfset result = {
-                ok: true,
-                doc_id: url.doc_id,
-                filename: doc.filename,
-                mime_type: doc.mime_type,
-                status: newStatus,
-                element_count: arrayLen(elements),
-                page_count: isDefined("pageCount") ? pageCount : 0,
-                errors: errors
-            }>
+            <cfset result = {ok: true, doc_id: url.doc_id, filename: doc.filename,
+                mime_type: doc.mime_type, status: newStatus,
+                element_count: arrayLen(elements), error_count: arrayLen(errors),
+                files: files}>
         </cfif>
     </cfif>
-
     <cfcatch>
         <cfset result = {ok: false, error: cfcatch.message, detail: cfcatch.detail}>
     </cfcatch>
 </cftry>
-
 <cfoutput>#serializeJSON(result)#</cfoutput>