Nail/contractor
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

v1.0.5: parser.cfm — PDF через PDFBox, TXT

Browse Source
This commit is contained in:
naeel
2026-06-18 16:15:00 +04:00
parent cf59d82d5b
commit 186c889513
2 changed files with 117 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
parser.cfm
+116
View File
@@ -0,0 +1,116 @@
<cfsetting showdebugoutput="no" enablecfoutputonly="true">
<cfheader name="Content-Type" value="application/json; charset=utf-8">
<cfparam name="url.doc_id" default="">
<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">
</cfquery>
<cfif doc.recordCount EQ 0>
<cfset result.error = "document not found">
<cfelse>
<cfset elements = []>
<cfset errors = []>
<!--- PDF --->
<cfif doc.mime_type EQ "application/pdf">
<cftry>
<!--- Base64 → байты --->
<cfset fileBytes = binaryDecode(doc.original_bytes, "base64")>
<!--- Байты → InputStream --->
<cfset bais = createObject("java", "java.io.ByteArrayInputStream").init(fileBytes)>
<!--- Открываем PDF --->
<cfset pdfDoc = createObject("java", "org.apache.pdfbox.pdmodel.PDDocument").load(bais)>
<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()>
<cfset bais.close()>
<cfcatch>
<cfset arrayAppend(errors, "PDFBox: " & cfcatch.message)>
</cfcatch>
</cftry>
<!--- DOCX/DOC — POI нет, пропускаем --->
<cfelseif doc.mime_type EQ "application/vnd.openxmlformats-officedocument.wordprocessingml.document"
OR doc.mime_type EQ "application/msword">
<cfset arrayAppend(errors, "DOCX/DOC parsing not available (POI not installed)")>
<!--- 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 line = trim(line)>
<cfif len(line)>
<cfset arrayAppend(elements, {type: "paragraph", style: "", text: line, page: 1})>
</cfif>
</cfloop>
<cfcatch>
<cfset arrayAppend(errors, "TXT: " & cfcatch.message)>
</cfcatch>
</cftry>
<cfelse>
<cfset arrayAppend(errors, "unsupported mime_type: " & doc.mime_type)>
</cfif>
<!--- Сохраняем результат --->
<cfset 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,
status = <cfqueryparam value="#newStatus#" cfsqltype="cf_sql_varchar">,
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
}>
</cfif>
</cfif>
<cfcatch>
<cfset result = {ok: false, error: cfcatch.message, detail: cfcatch.detail}>
</cfcatch>
</cftry>
<cfoutput>#serializeJSON(result)#</cfoutput>