@mdgate/hwp 0.5.0 → 0.5.2

package/dist/index.js

@@ -1,2 +1,2 @@
-import{CompoundFile as Fn,detectOleDoc as In,detectZipDoc as Xn,hasOleMagic as k,Package as kn}from"@mdgate/containers";import{ConvertError as p}from"@mdgate/core";import{documentToMarkdown as Mn}from"@mdgate/document";import{fileExtension as On}from"@mdgate/utils";import{hasOleMagic as Wn}from"@mdgate/containers";import{ConvertError as Nn}from"@mdgate/core";import{CompoundFile as Z,hasOleMagic as z,MAX_ENTRY_BYTES as V,MAX_RECORDS as Y}from"@mdgate/containers";import{ConvertError as s}from"@mdgate/core";import{emptyDocument as J,plain as Q}from"@mdgate/document";import{cleanText as C,collapseWs as D,InflateLimitError as j,inflateZlib as q,isAlphanumeric as K}from"@mdgate/utils";var x="HWP Document File",S=2,y=16,b=y+51,nn=new Set([1,2,3,4,5,6,7,8,11,12,13,14,15,16,17,21,22,23,26,27,28,29,30]);function d(n){if(n.length<x.length)return!1;for(let e=0;e<x.length;e+=1)if(n[e]!==x.charCodeAt(e))return!1;return!0}function B(n){if(z(n))return en(n);if(d(n))return L(un(n));throw s.malformed("not an HWP document")}function en(n){let e;try{e=Z.open(n)}catch(i){if(i instanceof s)throw i;throw s.malformed("not an OLE2 compound file")}if(e.exists("EncryptionInfo")||e.exists("EncryptedPackage"))throw s.encrypted();let t=A(e,"FileHeader"),r=t!==void 0?tn(t):void 0;if(r?.encrypted||r?.distribution)throw s.encrypted();let o=[],f=!1;for(let i=0;i<256;i+=1){let c=`BodyText/Section${i}`;if(!e.exists(c)){if(f||i>0)break;continue}f=!0;let m=W(e,c);if(m===void 0)continue;let u=cn(m,r?.compressed);rn(u,o)}if(o.length===0){let i=A(e,"PrvText");if(i!==void 0)an(i,o)}if(o.length===0)for(let i of e.readRootStorage()){if(i.type!==2)continue;let c=A(e,i.name);if(c!==void 0)g(c,o)}return L(o)}function tn(n){if(n.length<40)return;if(!d(n))return;let e=new DataView(n.buffer,n.byteOffset,n.byteLength).getUint32(36,!0);return{compressed:(e&1)!==0,encrypted:(e&2)!==0,distribution:(e&4)!==0}}function rn(n,e){let t=on(n);if(t.length>0){for(let r of t)e.push(r);return}g(n,e)}function on(n){if(n.length<4)return[];let e=new DataView(n.buffer,n.byteOffset,n.byteLength),t=[],r=0,o=0,f=0;while(r+4<=n.length&&o<Y){let i=e.getUint32(r,!0),c=i&1023,m=i>>>20&4095,u=m,l=r+4;if(m===4095){if(l+4>n.length)break;u=e.getUint32(l,!0),l+=4}if(u<0||l+u>n.length)break;if(o+=1,f=l+u,c===b){let a=fn(n.subarray(l,l+u));if(a.length>0)t.push(a)}r=l+u}if(t.length===0&&(o===0||f<n.length/2))return[];return t}function fn(n){let e=n.length&-2,t="",r=0;while(r+1<e){let o=n[r]|n[r+1]<<8;if(r+=2,o<32){if(nn.has(o))r+=14;else if(o===9)t+="\t";else if(o===10)t+=`
-`;else if(o===24||o===25)t+="-";else if(o===31)t+=" ";continue}if(o===65534||o===65535)continue;t+=String.fromCharCode(o)}return t}function cn(n,e){if(n.length>=2&&ln(n)){let t=H(n);if(t!==void 0)return t}if(e===!1)return n;if(e===!0){let t=H(n);if(t!==void 0)return t}return n}function H(n){try{return q(n,V)}catch(e){if(e instanceof j)throw s.resourceLimit("max_entry_bytes",e.message);if(e instanceof s&&e.isFatal())throw e;return}}function ln(n){if(n.length<2)return!1;let e=n[0],t=n[1];if(((e<<8)+t)%31!==0)return!1;return(e&15)===8}function un(n){let e=n.length>32?32:x.length,t=[];if(g(n.subarray(Math.min(e,n.length)),t),t.length===0)g(n,t);return t}function an(n,e){let t=n.length&-2,r=new TextDecoder("utf-16le").decode(n.subarray(0,t));for(let o of r.split(/\0+/)){let f=D(C(o)).trim();if(R(f))e.push(f)}}function g(n,e){let t="",r=()=>{if(t.length>=S)e.push(t);t=""},o=n.length&-2;for(let f=0;f<o;f+=2){let i=n[f]|n[f+1]<<8;if(i!==0&&sn(i))t+=String.fromCharCode(i);else r()}r()}function L(n){let e=new Set,t=J();for(let r of n){let o=D(C(r)).trim();if(!R(o)||e.has(o))continue;e.add(o),t.blocks.push({type:"paragraph",inlines:[Q(o)]})}if(t.blocks.length===0)throw s.malformed("no readable text in HWP document");return t}function sn(n){if(n===9||n===10||n===13||n===32)return!0;if(n<32||n>=127&&n<=159)return!1;return n!==65534&&n!==65535}function R(n){if(n.length<S)return!1;let e=0;for(let r of n)if(K(r))e+=1;if(e<2)return!1;let t=n.toLowerCase();if(t.includes("xmlns")||t.includes("http://schemas.")||t.includes("hwp document file"))return!1;return!0}function A(n,e){if(!n.exists(e))return;return W(n,e)}function W(n,e){try{return n.readStream(e)}catch(t){if(t instanceof s&&t.isFatal())throw t;return}}import{Package as dn}from"@mdgate/containers";import{ConvertError as h}from"@mdgate/core";import{cellSpanning as mn,emptyDocument as hn,GridBuilder as pn,inlinesAreEmpty as xn,plain as w,resolveHeaderRows as gn}from"@mdgate/document";import{cleanText as U}from"@mdgate/utils";function F(n){let e;try{e=dn.open(n)}catch(i){throw Rn(i)}let t=e.optionalXmlPart("Contents/header.xml")??e.optionalXmlPart("Contents/Header.xml"),r={styles:t!==void 0?Pn(t):new Map},o=wn(e);if(o.length===0)throw h.missingPart("Contents/section0.xml");let f=hn();for(let i of o){let c=e.optionalXmlPart(i);if(c===void 0)continue;P(c,f.blocks,r)}if(f.blocks.length===0)Dn(e,f);if(f.blocks.length===0)throw h.malformed("no readable text in HWPX document");return f}function wn(n){let e=[];for(let t of n.partNames())if(/^Contents\/section\d+\.xml$/i.test(t))e.push(t);return e.sort((t,r)=>N(t)-N(r)),e}function N(n){let e=n.match(/section(\d+)/i);return e!==null?Number(e[1]):0}function Pn(n){let e=new Map;for(let t of n.descendantsAny("style")){let r=t.attrAny("id");if(r===void 0)continue;let o=An(t);if(o!==void 0)e.set(r,o)}return e}function An(n){let e=(n.attrAny("type")??"").toLowerCase(),r=`${n.attrAny("name")??""} ${n.attrAny("engName")??""}`.match(/(?:개요|제목|outline|heading)\s*(\d+)/i);if(r!==null)return En(Number(r[1]));if(e==="outline"||e==="heading")return 1;return}function En(n){if(!Number.isFinite(n)||n<1)return 1;return n>6?6:n|0}function P(n,e,t){if(n.local==="tbl"){let r=I(n,t);if(r!==void 0)e.push(r);return}if(n.local==="p"){Un(n,e,t);return}for(let r of n.childElems())P(r,e,t)}function Un(n,e,t){let r=Tn(n,t),o=[],f=!1,i=()=>{if(xn(o)){o=[];return}if(r!==void 0&&!f)e.push({type:"heading",level:r,anchor:void 0,content:o}),f=!0;else e.push({type:"paragraph",inlines:o});o=[]},c=(m)=>{for(let u of m.children){if(u.type==="text"){let a=U(u.text);if(a.length>0)o.push(w(a));continue}let l=u.elem;switch(l.local){case"tbl":{i();let a=I(l,t);if(a!==void 0)e.push(a);break}case"t":{let a=U(l.text());if(a.length>0)o.push(w(a));break}case"lineBreak":case"line-break":o.push({type:"lineBreak"});break;case"tab":o.push(w("\t"));break;default:c(l)}}};c(n),i()}function Tn(n,e){let t=n.attrAny("styleIDRef")??n.attrAny("styleId")??n.attrAny("styleID");if(t===void 0)return;return e.styles.get(t)}function I(n,e){let t=vn(n);if(t.length===0)return;let r=new pn;for(let f of t){r.nextRow();for(let i of Hn(f))r.place(mn(Cn(i,e),_(i,"colSpan"),_(i,"rowSpan")))}let o=r.finish("data");if(o.grid.length===0)return;return o.headerRows=gn(o,0),{type:"table",table:o}}function vn(n){let e=n.childElems().filter((t)=>t.local==="tr");if(e.length>0)return e;return[...n.descendantsAny("tr")]}function Hn(n){let e=n.childElems().filter((t)=>t.local==="tc");if(e.length>0)return e;return[...n.descendantsAny("tc")]}function Cn(n,e){let t=[],r=n.childElems().find((o)=>o.local==="subList");if(r!==void 0)return P(r,t,e),t;for(let o of n.childElems())if(o.local==="p"||o.local==="tbl")P(o,t,e);return t}function _(n,e){let t=n.attrAny(e)??n.childElems().find((o)=>o.local==="cellSpan")?.attrAny(e);if(t===void 0||t.length===0)return 1;let r=Number(t);if(!Number.isFinite(r)||r<1)return 1;return r>1000?1000:r|0}function Dn(n,e){let t=n.optionalPart("Preview/PrvText.txt")??n.optionalPart("Preview/PrvText.dat")??n.optionalPart("Preview/PrvText");if(t===void 0)return;for(let r of Ln(Sn(t)))e.blocks.push({type:"paragraph",inlines:[w(r)]})}function Sn(n){if(n.length>=2&&n[0]===255&&n[1]===254)return new TextDecoder("utf-16le").decode(n);if(n.length>=2&&n[0]===254&&n[1]===255)return new TextDecoder("utf-16be").decode(n);let e=n.length&-2;if(e>=4&&Bn(n.subarray(0,e)))return new TextDecoder("utf-16le").decode(n.subarray(0,e));return new TextDecoder("utf-8",{fatal:!1}).decode(n)}function Bn(n){let e=0;for(let t=1;t<n.length;t+=2)if(n[t]===0)e+=1;return e*2>=n.length/2}function Ln(n){let e=[];for(let t of n.split(/\0+/)){let r=U(t).trim();if(r.length>0)e.push(r)}return e}function Rn(n){if(n instanceof h){if(n.code==="encrypted"||E(n.message)||E(n.detail??""))return h.encrypted();return n}let e=n instanceof Error?n.message:String(n);if(E(e))return h.encrypted();return h.malformed(`not a readable zip archive: ${e}`)}function E(n){let e=n.toLowerCase();return e.includes("encrypted")||e.includes("password")}function X(n){if(_n(n))return F(n);if(Wn(n)||d(n))return B(n);throw Nn.malformed("not an HWP document")}function _n(n){return n.length>=4&&n[0]===80&&n[1]===75&&n[2]===3&&n[3]===4}var Gn=new Set(["hwp","hwpx","hwt","hwtx"]);function v(){return{id:"hwp",sniff(n,e){if(O(n)==="hwpx"||M(n))return 2;if(G(n)==="hwp"||Zn(n))return 2;if(d(n))return 2;if(e?.path!==void 0&&Gn.has(On(e.path)??""))return 1;return 0},convert(n){return $n(n),{markdown:Mn(X(n))}}}}function $n(n){if(Vn(n))throw p.unsupported("pdf");if($(n)){let e=O(n);if(e==="hwpx")return;if(e!==void 0)throw p.unsupported(e);if(M(n))return;throw p.unsupported("zip")}if(k(n)){let e=G(n);if(e==="hwp")return;if(e!==void 0)throw p.unsupported(e);return}if(d(n))return;throw p.unsupported("hwp")}function M(n){if(!$(n))return!1;let e;try{e=kn.open(n)}catch{return!1}for(let r of e.partNames()){let o=r.toLowerCase();if(o==="contents/content.hpf")return!0;if(/^contents\/section\d+\.xml$/.test(o))return!0}let t=e.optionalPart("mimetype");if(t!==void 0){let o=new TextDecoder("utf-8",{fatal:!1}).decode(t).trim().toLowerCase().replace(/-template$/,"");if(o==="application/hwp+zip"||o==="application/vnd.hancom.hwpx")return!0}return!1}function Zn(n){if(!k(n))return!1;try{let e=Fn.open(n);for(let t of e.readRootStorage())if(zn(t.name))return!0}catch{return!1}return!1}function zn(n){let e=n.charCodeAt(0)===5?n.slice(1):n;return T(n,"FileHeader")||T(e,"HwpSummaryInformation")||T(n,"HWPDocumentInfo")}function T(n,e){return n.length===e.length&&n.toLowerCase()===e.toLowerCase()}function O(n){return Xn(n)}function G(n){return In(n)}function $(n){return n.length>=4&&n[0]===80&&n[1]===75&&n[2]===3&&n[3]===4}function Vn(n){let e=Yn(n);return Jn(n,"%PDF-",e)}function Yn(n){let e=0;if(n.length>=3&&n[0]===239&&n[1]===187&&n[2]===191)e=3;while(e<n.length){let t=n[e];if(t!==9&&t!==10&&t!==13&&t!==32)break;e+=1}return e}function Jn(n,e,t){if(t+e.length>n.length)return!1;for(let r=0;r<e.length;r+=1)if(n[t+r]!==e.charCodeAt(r))return!1;return!0}import{create as Qn}from"@mdgate/core";var jn=Qn([v()]);function qn(n,e){return jn(n,e)}export{qn as toMarkdown,v as hwp};
+import{CompoundFile as Kn,detectOleDoc as Qn,detectZipDoc as Yn,hasOleMagic as Z,Package as jn}from"@mdgate/containers";import{ConvertError as p}from"@mdgate/core";import{documentToMarkdown as qn}from"@mdgate/document";import{fileExtension as n0}from"@mdgate/utils";import{hasOleMagic as Zn}from"@mdgate/containers";import{ConvertError as zn}from"@mdgate/core";import{CompoundFile as tn,hasOleMagic as xn}from"@mdgate/containers";import{ConvertError as m}from"@mdgate/core";import{emptyDocument as rn,plain as on}from"@mdgate/document";import{cleanText as R,collapseWs as I,inflateRaw as fn,inflateZlib as cn,isAlphanumeric as an}from"@mdgate/utils";var g=Uint8Array.from([99,124,119,123,242,107,111,197,48,1,103,43,254,215,171,118,202,130,201,125,250,89,71,240,173,212,162,175,156,164,114,192,183,253,147,38,54,63,247,204,52,165,229,241,113,216,49,21,4,199,35,195,24,150,5,154,7,18,128,226,235,39,178,117,9,131,44,26,27,110,90,160,82,59,214,179,41,227,47,132,83,209,0,237,32,252,177,91,106,203,190,57,74,76,88,207,208,239,170,251,67,77,51,133,69,249,2,127,80,60,159,168,81,163,64,143,146,157,56,245,188,182,218,33,16,255,243,210,205,12,19,236,95,151,68,23,196,167,126,61,100,93,25,115,96,129,79,220,34,42,144,136,70,238,184,20,222,94,11,219,224,50,58,10,73,6,36,92,194,211,172,98,145,149,228,121,231,200,55,109,141,213,78,169,108,86,244,234,101,122,174,8,186,120,37,46,28,166,180,198,232,221,116,31,75,189,139,138,112,62,181,102,72,3,246,14,97,53,87,185,134,193,29,158,225,248,152,17,105,217,142,148,155,30,135,233,206,85,40,223,140,161,137,13,191,230,66,104,65,153,45,15,176,84,187,22]),Y=Uint8Array.from([82,9,106,213,48,54,165,56,191,64,163,158,129,243,215,251,124,227,57,130,155,47,255,135,52,142,67,68,196,222,233,203,84,123,148,50,166,194,35,61,238,76,149,11,66,250,195,78,8,46,161,102,40,217,36,178,118,91,162,73,109,139,209,37,114,248,246,100,134,104,152,22,212,164,92,204,93,101,182,146,108,112,72,80,253,237,185,218,94,21,70,87,167,141,157,132,144,216,171,0,140,188,211,10,247,228,88,5,184,179,69,6,208,44,30,143,202,63,15,2,193,175,189,3,1,19,138,107,58,145,17,65,79,103,220,234,151,242,207,206,240,180,230,115,150,172,116,34,231,173,53,133,226,249,55,232,28,117,223,110,71,241,26,113,29,41,197,137,111,183,98,14,170,24,190,27,252,86,62,75,198,210,121,32,154,219,192,254,120,205,90,244,31,221,168,51,136,7,199,49,177,18,16,89,39,128,236,95,96,81,127,169,25,181,74,13,45,229,122,159,147,201,156,239,160,224,59,77,174,42,245,176,200,235,187,60,131,83,153,97,23,43,4,126,186,119,214,38,225,105,20,99,85,33,12,125]),j=Uint8Array.from([0,1,2,4,8,16,32,64,128,27,54]);function H(n,e){let t=n.length-n.length%16,r=q(e),x=new Uint8Array(t),o=new Uint8Array(16);for(let i=0;i<t;i+=16)o.set(n.subarray(i,i+16)),nn(o,r),x.set(o,i);return x}function q(n){let e=new Uint8Array(176);e.set(n.subarray(0,16));for(let t=16,r=1;t<176;t+=4){let x=e[t-4],o=e[t-3],i=e[t-2],f=e[t-1];if(t%16===0){let c=x;x=(g[o]^j[r++])&255,o=g[i],i=g[f],f=g[c]}e[t]=e[t-16]^x,e[t+1]=e[t-15]^o,e[t+2]=e[t-14]^i,e[t+3]=e[t-13]^f}return e}function nn(n,e){P(n,e,10);for(let t=9;t>=1;t-=1)D(n),C(n),P(n,e,t),en(n);D(n),C(n),P(n,e,0)}function P(n,e,t){let r=t*16;for(let x=0;x<16;x+=1)n[x]^=e[r+x]}function C(n){for(let e=0;e<16;e+=1)n[e]=Y[n[e]]}function D(n){let e=n[13];n[13]=n[9],n[9]=n[5],n[5]=n[1],n[1]=e,e=n[2],n[2]=n[10],n[10]=e,e=n[6],n[6]=n[14],n[14]=e,e=n[3],n[3]=n[7],n[7]=n[11],n[11]=n[15],n[15]=e}function en(n){for(let e=0;e<4;e+=1){let t=e*4,r=n[t],x=n[t+1],o=n[t+2],i=n[t+3];n[t]=a(r,14)^a(x,11)^a(o,13)^a(i,9),n[t+1]=a(r,9)^a(x,14)^a(o,11)^a(i,13),n[t+2]=a(r,13)^a(x,9)^a(o,14)^a(i,11),n[t+3]=a(r,11)^a(x,13)^a(o,9)^a(i,14)}}function a(n,e){let t=0;for(let r=0;r<8;r+=1){if((e&1)!==0)t^=n;let x=(n&128)!==0;if(n=n<<1&255,x)n^=27;e>>=1}return t}var A="HWP Document File",F=2,un=16,ln=un+51,dn=new Set([1,2,3,4,5,6,7,8,11,12,13,14,15,16,17,21,22,23,26,27,28,29,30]);function s(n){if(n.length<A.length)return!1;for(let e=0;e<A.length;e+=1)if(n[e]!==A.charCodeAt(e))return!1;return!0}function W(n){if(xn(n))return sn(n);if(s(n))return L(En(n));throw m.malformed("not an HWP document")}function sn(n){let e;try{e=tn.open(n)}catch(o){if(o instanceof m)throw o;throw m.malformed("not an OLE2 compound file")}if(e.exists("EncryptionInfo")||e.exists("EncryptedPackage"))throw m.encrypted();let t=E(e,"FileHeader"),r=t!==void 0?An(t):void 0;if(r?.encrypted)throw m.encrypted();let x=[];if(r?.distribution)B(e,"ViewText",r.compressed,x,!0);else B(e,"BodyText",r?.compressed,x,!1);if(x.length===0){let o=E(e,"PrvText");if(o!==void 0)yn(o,x)}if(x.length===0)for(let o of e.readRootStorage()){if(o.type!==2)continue;let i=E(e,o.name);if(i!==void 0)w(i,x)}return L(x)}function B(n,e,t,r,x){let o=!1;for(let i=0;;i+=1){let f=`${e}/Section${i}`;if(!n.exists(f)){if(o||i>0)break;continue}o=!0;let c=X(n,f);if(c===void 0)continue;let l=x?mn(c):c;if(l===void 0)continue;let u=x?pn(l,t):k(l,t);wn(u,r)}}function mn(n){if(n.length<260)return;let e=hn(n.subarray(4,260));if(e===void 0)return;return H(n.subarray(260),e)}function hn(n){if(n.length<256)return;let e=n.subarray(0,256).slice(),t=new DataView(e.buffer,e.byteOffset,4).getInt32(0,!0),r=0,x=0,o=t|0,i=()=>{return o=Math.imul(o,214013)+2531011|0,o>>>16&32767};for(let c=0;c<256;c+=1,x-=1){if(x===0)r=i()&255,x=(i()&15)+1;if(c>=4)e[c]=e[c]^r}let f=4+(e[0]&15);if(f+16>e.length)return;return e.subarray(f,f+16)}function pn(n,e){if(e!==!0)return n;let t=gn(n);if(t!==void 0)return t;return k(n,!0)}function gn(n){if(n.length===0)return;try{return fn(n,Number.MAX_SAFE_INTEGER)}catch{return}}function An(n){if(n.length<40)return;if(!s(n))return;let e=new DataView(n.buffer,n.byteOffset,n.byteLength).getUint32(36,!0);return{compressed:(e&1)!==0,encrypted:(e&2)!==0,distribution:(e&4)!==0}}function wn(n,e){let t=Un(n);if(t.length>0){for(let r of t)e.push(r);return}w(n,e)}function Un(n){if(n.length<4)return[];let e=new DataView(n.buffer,n.byteOffset,n.byteLength),t=[],r=0,x=0,o=0;while(r+4<=n.length){let i=e.getUint32(r,!0),f=i&1023,c=i>>>20&4095,l=c,u=r+4;if(c===4095){if(u+4>n.length)break;l=e.getUint32(u,!0),u+=4}if(l<0||u+l>n.length)break;if(x+=1,o=u+l,f===ln){let d=bn(n.subarray(u,u+l));if(d.length>0)t.push(d)}r=u+l}if(t.length===0&&(x===0||o<n.length/2))return[];return t}function bn(n){let e=n.length&-2,t="",r=0;while(r+1<e){let x=n[r]|n[r+1]<<8;if(r+=2,x<32){if(dn.has(x))r+=14;else if(x===9)t+="\t";else if(x===10)t+=`
+`;else if(x===24||x===25)t+="-";else if(x===31)t+=" ";continue}if(x===65534||x===65535)continue;t+=String.fromCharCode(x)}return t}function k(n,e){if(n.length>=2&&Pn(n)){let t=N(n);if(t!==void 0)return t}if(e===!1)return n;if(e===!0){let t=N(n);if(t!==void 0)return t}return n}function N(n){try{return cn(n,Number.MAX_SAFE_INTEGER)}catch{return}}function Pn(n){if(n.length<2)return!1;let e=n[0],t=n[1];if(((e<<8)+t)%31!==0)return!1;return(e&15)===8}function En(n){let e=n.length>32?32:A.length,t=[];if(w(n.subarray(Math.min(e,n.length)),t),t.length===0)w(n,t);return t}function yn(n,e){let t=n.length&-2,r=new TextDecoder("utf-16le").decode(n.subarray(0,t));for(let x of r.split(/\0+/)){let o=I(R(x)).trim();if(O(o))e.push(o)}}function w(n,e){let t="",r=()=>{if(t.length>=F)e.push(t);t=""},x=n.length&-2;for(let o=0;o<x;o+=2){let i=n[o]|n[o+1]<<8;if(i!==0&&vn(i))t+=String.fromCharCode(i);else r()}r()}function L(n){let e=new Set,t=rn();for(let r of n){let x=I(R(r)).trim();if(!O(x)||e.has(x))continue;e.add(x),t.blocks.push({type:"paragraph",inlines:[on(x)]})}if(t.blocks.length===0)throw m.malformed("no readable text in HWP document");return t}function vn(n){if(n===9||n===10||n===13||n===32)return!0;if(n<32||n>=127&&n<=159)return!1;return n!==65534&&n!==65535}function O(n){if(n.length<F)return!1;let e=0;for(let r of n)if(an(r))e+=1;if(e<2)return!1;let t=n.toLowerCase();if(t.includes("xmlns")||t.includes("http://schemas.")||t.includes("hwp document file"))return!1;return!0}function E(n,e){if(!n.exists(e))return;return X(n,e)}function X(n,e){try{return n.readStream(e)}catch(t){if(t instanceof m&&t.isFatal())throw t;return}}import{Package as Tn}from"@mdgate/containers";import{ConvertError as h}from"@mdgate/core";import{cellSpanning as Sn,emptyDocument as Cn,GridBuilder as Dn,inlinesAreEmpty as Hn,plain as U,resolveHeaderRows as Bn}from"@mdgate/document";import{cleanText as v}from"@mdgate/utils";function G(n){let e;try{e=Tn.open(n)}catch(i){throw $n(i)}let t=e.optionalXmlPart("Contents/header.xml")??e.optionalXmlPart("Contents/Header.xml"),r={styles:t!==void 0?Rn(t):new Map},x=Nn(e);if(x.length===0)throw h.missingPart("Contents/section0.xml");let o=Cn();for(let i of x){let f=e.optionalXmlPart(i);if(f===void 0)continue;b(f,o.blocks,r)}if(o.blocks.length===0)_n(e,o);if(o.blocks.length===0)throw h.malformed("no readable text in HWPX document");return o}function Nn(n){let e=[];for(let t of n.partNames())if(/^Contents\/section\d+\.xml$/i.test(t))e.push(t);return e.sort((t,r)=>_(t)-_(r)),e}function _(n){let e=n.match(/section(\d+)/i);return e!==null?Number(e[1]):0}function Rn(n){let e=new Map;for(let t of n.descendantsAny("style")){let r=t.attrAny("id");if(r===void 0)continue;let x=In(t);if(x!==void 0)e.set(r,x)}return e}function In(n){let e=(n.attrAny("type")??"").toLowerCase(),r=`${n.attrAny("name")??""} ${n.attrAny("engName")??""}`.match(/(?:개요|제목|outline|heading)\s*(\d+)/i);if(r!==null)return Fn(Number(r[1]));if(e==="outline"||e==="heading")return 1;return}function Fn(n){if(!Number.isFinite(n)||n<1)return 1;return n>6?6:n|0}function b(n,e,t){if(n.local==="tbl"){let r=V(n,t);if(r!==void 0)e.push(r);return}if(n.local==="p"){Wn(n,e,t);return}for(let r of n.childElems())b(r,e,t)}function Wn(n,e,t){let r=kn(n,t),x=[],o=!1,i=()=>{if(Hn(x)){x=[];return}if(r!==void 0&&!o)e.push({type:"heading",level:r,anchor:void 0,content:x}),o=!0;else e.push({type:"paragraph",inlines:x});x=[]},f=(c)=>{for(let l of c.children){if(l.type==="text"){let d=v(l.text);if(d.length>0)x.push(U(d));continue}let u=l.elem;switch(u.local){case"tbl":{i();let d=V(u,t);if(d!==void 0)e.push(d);break}case"t":{let d=v(u.text());if(d.length>0)x.push(U(d));break}case"lineBreak":case"line-break":x.push({type:"lineBreak"});break;case"tab":x.push(U("\t"));break;default:f(u)}}};f(n),i()}function kn(n,e){let t=n.attrAny("styleIDRef")??n.attrAny("styleId")??n.attrAny("styleID");if(t===void 0)return;return e.styles.get(t)}function V(n,e){let t=Ln(n);if(t.length===0)return;let r=new Dn;for(let o of t){r.nextRow();for(let i of On(o))r.place(Sn(Xn(i,e),M(i,"colSpan"),M(i,"rowSpan")))}let x=r.finish("data");if(x.grid.length===0)return;return x.headerRows=Bn(x,0),{type:"table",table:x}}function Ln(n){let e=n.childElems().filter((t)=>t.local==="tr");if(e.length>0)return e;return[...n.descendantsAny("tr")]}function On(n){let e=n.childElems().filter((t)=>t.local==="tc");if(e.length>0)return e;return[...n.descendantsAny("tc")]}function Xn(n,e){let t=[],r=n.childElems().find((x)=>x.local==="subList");if(r!==void 0)return b(r,t,e),t;for(let x of n.childElems())if(x.local==="p"||x.local==="tbl")b(x,t,e);return t}function M(n,e){let t=n.attrAny(e)??n.childElems().find((x)=>x.local==="cellSpan")?.attrAny(e);if(t===void 0||t.length===0)return 1;let r=Number(t);if(!Number.isFinite(r)||r<1)return 1;return r>1000?1000:r|0}function _n(n,e){let t=n.optionalPart("Preview/PrvText.txt")??n.optionalPart("Preview/PrvText.dat")??n.optionalPart("Preview/PrvText");if(t===void 0)return;for(let r of Vn(Mn(t)))e.blocks.push({type:"paragraph",inlines:[U(r)]})}function Mn(n){if(n.length>=2&&n[0]===255&&n[1]===254)return new TextDecoder("utf-16le").decode(n);if(n.length>=2&&n[0]===254&&n[1]===255)return new TextDecoder("utf-16be").decode(n);let e=n.length&-2;if(e>=4&&Gn(n.subarray(0,e)))return new TextDecoder("utf-16le").decode(n.subarray(0,e));return new TextDecoder("utf-8",{fatal:!1}).decode(n)}function Gn(n){let e=0;for(let t=1;t<n.length;t+=2)if(n[t]===0)e+=1;return e*2>=n.length/2}function Vn(n){let e=[];for(let t of n.split(/\0+/)){let r=v(t).trim();if(r.length>0)e.push(r)}return e}function $n(n){if(n instanceof h){if(n.code==="encrypted"||y(n.message)||y(n.detail??""))return h.encrypted();return n}let e=n instanceof Error?n.message:String(n);if(y(e))return h.encrypted();return h.malformed(`not a readable zip archive: ${e}`)}function y(n){let e=n.toLowerCase();return e.includes("encrypted")||e.includes("password")}function $(n){if(Jn(n))return G(n);if(Zn(n)||s(n))return W(n);throw zn.malformed("not an HWP document")}function Jn(n){return n.length>=4&&n[0]===80&&n[1]===75&&n[2]===3&&n[3]===4}var e0=new Set(["hwp","hwpx","hwt","hwtx"]);function S(){return{id:"hwp",sniff(n,e){if(J(n)==="hwpx"||z(n))return 2;if(K(n)==="hwp"||x0(n))return 2;if(s(n))return 2;if(e?.path!==void 0&&e0.has(n0(e.path)??""))return 1;return 0},convert(n){return t0(n),{markdown:qn($(n))}}}}function t0(n){if(o0(n))throw p.unsupported("pdf");if(Q(n)){let e=J(n);if(e==="hwpx")return;if(e!==void 0)throw p.unsupported(e);if(z(n))return;throw p.unsupported("zip")}if(Z(n)){let e=K(n);if(e==="hwp")return;if(e!==void 0)throw p.unsupported(e);return}if(s(n))return;throw p.unsupported("hwp")}function z(n){if(!Q(n))return!1;let e;try{e=jn.open(n)}catch{return!1}for(let r of e.partNames()){let x=r.toLowerCase();if(x==="contents/content.hpf")return!0;if(/^contents\/section\d+\.xml$/.test(x))return!0}let t=e.optionalPart("mimetype");if(t!==void 0){let x=new TextDecoder("utf-8",{fatal:!1}).decode(t).trim().toLowerCase().replace(/-template$/,"");if(x==="application/hwp+zip"||x==="application/vnd.hancom.hwpx")return!0}return!1}function x0(n){if(!Z(n))return!1;try{let e=Kn.open(n);for(let t of e.readRootStorage())if(r0(t.name))return!0}catch{return!1}return!1}function r0(n){let e=n.charCodeAt(0)===5?n.slice(1):n;return T(n,"FileHeader")||T(e,"HwpSummaryInformation")||T(n,"HWPDocumentInfo")}function T(n,e){return n.length===e.length&&n.toLowerCase()===e.toLowerCase()}function J(n){return Yn(n)}function K(n){return Qn(n)}function Q(n){return n.length>=4&&n[0]===80&&n[1]===75&&n[2]===3&&n[3]===4}function o0(n){let e=i0(n);return f0(n,"%PDF-",e)}function i0(n){let e=0;if(n.length>=3&&n[0]===239&&n[1]===187&&n[2]===191)e=3;while(e<n.length){let t=n[e];if(t!==9&&t!==10&&t!==13&&t!==32)break;e+=1}return e}function f0(n,e,t){if(t+e.length>n.length)return!1;for(let r=0;r<e.length;r+=1)if(n[t+r]!==e.charCodeAt(r))return!1;return!0}import{create as c0}from"@mdgate/core";var a0=c0([S()]);function u0(n,e){return a0(n,e)}export{u0 as toMarkdown,S as hwp};

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@mdgate/hwp",
-  "version": "0.5.0",
+  "version": "0.5.2",
   "description": "mdgate Hangul HWP / HWPX converter",
   "type": "module",
   "license": "MIT",
@@ -22,10 +22,10 @@
     "prepublishOnly": "bun run build"
   },
   "dependencies": {
-    "@mdgate/containers": "0.5.0",
-    "@mdgate/core": "0.5.0",
-    "@mdgate/document": "0.5.0",
-    "@mdgate/utils": "0.5.0"
+    "@mdgate/containers": "0.5.2",
+    "@mdgate/core": "0.5.2",
+    "@mdgate/document": "0.5.2",
+    "@mdgate/utils": "0.5.2"
   },
   "publishConfig": {
     "access": "public"