完整源码：
链接: https://pan.baidu.com/s/1MglwXjH-59Nbz2ntWfGmvQ 提取码: hytu
这是本人（liigo）独立实现的SGF格式围棋棋谱文件解析器，本文介绍其实现细节。网络上肯定可以找到完善的开源的SGF解析器，这是毋庸置疑的，我不直接使用它们，也不参考它们的实现代码，而是自己独立编码实现，是有原因的，因为我想自己重复发明轮子，并且认为这样更有助于提高我的编码能力。（关于我的“一定要学会重复发明轮子”的不成熟的论调，今后我将会专门撰文表述。）
我（liigo）开发的这个SGF解析器，采用基于事件的简单API，类似于XML解析器中的SAX(Simple API for XML)。这种解析器的核心是：由用户事先提供一系列回调函数，解析器在解析的过程中，依次调用相关的回调函数并传入相应参数，用户程序在回调函数中做出相应的处理。此类解析器属于轻量级的解析器，解析速度快，占用内存少，结构清晰易于实现，只是相对来说不如基于DOM的解析器方便使用。
SGF格式，Smart Game Format，被设计用来记录多种游戏类棋谱的通用格式，在围棋领域被发扬光大，是用于描述围棋棋谱的最重要也最通用的形式。它是纯文本的、基于树（TREE）的结构，便于识别、存储和传输。其格式简洁实用，也非常易于编程解析。SGF格式官方规范网址为：http://www.red-bean.com/sgf/。（说到围棋棋谱，不得不赞叹一下，它只需用一幅图就可以完整还原一盘棋从始至终的风云变幻；作为对比，象棋一幅图只能描述对弈中某一时刻的场景。）
SGF的主要结构由树（GameTree）、节点序列（Sequence）、节点（Node）、属性（Property）等组成。其中“属性”为最重要的基本单位，它由属性标识（PropIdent）和属性值（PropValue）组成。由分号“;”分隔的多个属性，称为节点。多个节点顺序排列称为节点序列。由括号“(”“)”括起来的节点序列，称为树，树中可包含子树。SGF的EBNF定义如下
（参见http://www.red-bean.com/sgf/sgf4.html#ebnf-def）：
Collection = GameTree { GameTree }
GameTree = "(" Sequence { GameTree } ")"
Sequence = Node { Node }
Node = ";" { Property }
Property = PropIdent PropValue { PropValue }
PropIdent = UcLetter { UcLetter }
PropValue = "[" CValueType "]"
CValueType = (ValueType | Compose)
ValueType = (None | Number | Real | Double | Color | SimpleText | Text | Point | Move | Stone)
以下是一个简单的有一定代表性的SGF文本，先让大家有一个感性认识：
(;FF[4]GM[1]SZ[19]FG[257:Figure 1]PM[1] PB[Takemiya Masaki]BR[9 dan]PW[Cho Chikun] WR[9 dan]RE[W+Resign]KM[5.5]TM[28800]DT[1996-10-18,19] EV[21st Meijin]RO[2 (final)]SO[Go World #78]US[Arno Hollosi] ;B[pd];W[dp];B[pp];W[dd];B[pj];W[nc];B[oe];W[qc];B[pc];W[qd] (;B[qf];W[rf];B[rg];W[re];B[qg];W[pb];B[ob];W[qb] (;B[mp];W[fq];B[ci];W[cg];B[dl];W[cn];B[qo];W[ec];B[jp];W[jd] ;B[ei];W[eg];B[kk]LB[qq:a][dj:b][ck:c][qp:d]N[Figure 1] ;W[me]FG[257:Figure 2];B[kf];W[ke];B[lf];W[jf];B[jg] (;W[mf];B[if];W[je];B[ig];W[mg];B[mj];W[mq];B[lq];W[nq] (;B[lr];W[qq];B[pq];W[pr];B[rq];W[rr];B[rp];W[oq];B[mr];W[oo];B[mn] (;W[nr];B[qp]LB[kd:a][kh:b]N[Figure 2] ;W[pk]FG[257:Figure 3];B[pm];W[oj];B[ok];W[qr];B[os];W[ol];B[nk];W[qj] ;B[pi];W[pl];B[qm];W[ns];B[sr];W[om];B[op];W[qi];B[oi] (;W[rl];B[qh];W[rm];B[rn];W[ri];B[ql];W[qk];B[sm];W[sk];B[sh];W[og] ;B[oh];W[np];B[no];W[mm];B[nn];W[lp];B[kp];W[lo];B[ln];W[ko];B[mo] ;W[jo];B[km]N[Figure 3]) (;W[ql]VW[ja:ss]FG[257:Dia. 6]MN[1];B[rm];W[ph];B[oh];W[pg];B[og];W[pf] ;B[qh];W[qe];B[sh];W[of];B[sj]TR[oe][pd][pc][ob]LB[pe:a][sg:b][si:c] N[Diagram 6])) (;W[no]VW[jj:ss]FG[257:Dia. 5]MN[1];B[pn]N[Diagram 5])) (;B[pr]FG[257:Dia. 4]MN[1];W[kq];B[lp];W[lr];B[jq];W[jr];B[kp];W[kr];B[ir] ;W[hr]LB[is:a][js:b][or:c]N[Diagram 4])) (;W[if]FG[257:Dia. 3]MN[1];B[mf];W[ig];B[jh]LB[ki:a]N[Diagram 3])) (;W[oc]VW[aa:sk]FG[257:Dia. 2]MN[1];B[md];W[mc];B[ld]N[Diagram 2])) (;B[qe]VW[aa:sj]FG[257:Dia. 1]MN[1];W[re];B[qf];W[rf];B[qg];W[pb];B[ob] ;W[qb]LB[rg:a]N[Diagram 1]))
熟悉编写文本解析器的程序员朋友应该都清楚，根据EBNF定义，编写对应的解析器，是相当简单和直观的，貌似只是一项翻译性的工作。本人实现SGF解析器，再次印证了这个观点，大部分情况下，我只是按部就班地将EBNF翻译为C语言代码而已，呵呵。
我首先设计了“SGFParseContext”结构，用于保存解析器工作期间的相关数据：
typedef struct _tagSGFParseContext
{
void* pUserData;
int treeIndex;
PFN_ON_TREE pfnOnTree;
PFN_ON_TREE_END pfnOnTreeEnd;
PFN_ON_NODE pfnOnNode;
PFN_ON_NODE_END pfnOnNodeEnd;
PFN_ON_PROPERTY pfnOnProperty;
char idBuffer[16];
char* valueBuffer;
int valueBufferSize;
}
SGFParseContext;
相应的还有初始化和清理SGFParseContext结构的函数，initSGFParseContext, cleanupSGFParseContext，皆不是本解析器的关键，略过不提。
接着我(liigo)设计了五个回调函数的函数原形：
typedef void (*PFN_ON_TREE) (SGFParseContext* pContext, const char* szTreeHeader, int treeIndex); typedef void (*PFN_ON_TREE_END) (SGFParseContext* pContext, int treeIndex); typedef void (*PFN_ON_NODE) (SGFParseContext* pContext, const char* szNodeHeader); typedef void (*PFN_ON_NODE_END) (SGFParseContext* pContext); typedef void (*PFN_ON_PROPERTY) (SGFParseContext* pContext, const char* szID, const char* szValue);
这五个回调函数，将分别在解析器解析到“树开始”“树结束”“节点开始”“节点结束”“遇到属性”时，由解析器调用。解析器调用每个回调函数时，都会传入必需的参数，供回调函数即时取用。
下面正式开始解析工作。整个解析器被分为 parseProperty, parseNode, parseNodeSequence, parseGameTree, parseSGF 几大部分顺序解析，属于至底向上的分析实现模式。这几大部分，也分别对应着SGF的EBNF定义中的某一项。所有解析函数都接收参数 const char* szCollection, int fromPos，之前的解析函数将决定后续解析函数的起始解析位置。
第一步，解析属性（parseProperty）。此处关键的是要定位到属性值（szValue）开始和结束符号“[”和“]，两者之间的是属性值，“[”之前的则是属性标识（szID）。由于[和]之间可能存在转义字符“/”，不能简单地搜索字符“]”，必须花相当篇幅的代码处理转义字符（我用局部变量in_escape记录转义状态并进行分别处理）。此外要为提取出的属性标识和属性值分配足够的存储空间，以便传递到用户回调函数，前者不会太长使用静态分配，后者变长则使用动态分配（同时自动预分配存储空间，缓存，避免频繁申请内存）。代码如下：
//Property: id[value]
int parseProperty(SGFParseContext* pContext, const char* szCollection, int fromPos)
{
const char* szFromPos;
int lindex;
int nIDBufferSize = sizeof(pContext->idBuffer) - 1;
assert(szCollection && fromPos >= 0);
szFromPos = szCollection + fromPos;
lindex = findchar(szFromPos, -1, '[');
assert(lindex > 0 && lindex < nIDBufferSize);
if(lindex > 0 && lindex < nIDBufferSize)
{
memcpy(pContext->idBuffer, szFromPos, lindex);
pContext->idBuffer[lindex] = '/0';
if(isTextPropertyID(pContext->idBuffer))
{
//parse the text or simple-text value, consider the '/' escape character
const char* s = szFromPos + lindex + 1;
char c;
int in_escape = 0;
int valuelen = 0;
getEnoughBuffer(pContext, 1024);
pContext->valueBuffer[0] = '/0';
while(1)
{
c = *s;
assert(c);
if(!in_escape)
{
if(c == '//')
{
in_escape = 1;
}
else if(c == ']')
{
break;
}
else
{
getEnoughBuffer(pContext, valuelen + 1);
pContext->valueBuffer[valuelen++] = c;
}
}
else
{
//ignore the newline after '/'
if(c != '/r' && c != '/n')
{
getEnoughBuffer(pContext, valuelen + 1);
pContext->valueBuffer[valuelen++] = c;
}
else
{
char nc = *(s+1);
if(nc)
{
if((c=='/r' && nc=='/n') || (c=='/n' && nc=='/r'))
s++;
}
}
in_escape = 0;
}
s++;
}
getEnoughBuffer(pContext, valuelen + 1);
pContext->valueBuffer[valuelen] = '/0';
if(pContext->pfnOnProperty)
pContext->pfnOnProperty(pContext, pContext->idBuffer, pContext->valueBuffer);
return (s - szCollection + 1);
}
else
{
int rindex = findchar(szFromPos, -1, ']');
int nNeedBufferSize = rindex - lindex - 1;
assert(rindex >= 0);
getEnoughBuffer(pContext, nNeedBufferSize);
memcpy(pContext->valueBuffer, szFromPos + lindex + 1, nNeedBufferSize);
pContext->valueBuffer[nNeedBufferSize] = '/0';
if(pContext->pfnOnProperty)
pContext->pfnOnProperty(pContext, pContext->idBuffer, pContext->valueBuffer);
return (fromPos + rindex + 1);
}
}
return -1;
}
第二步，解析节点（parseNode）。分号“;”跟后面N个属性，一个while循环调用parseProperty()逐个解析属性即可：
//Node: ; {property}
int parseNode(SGFParseContext* pContext, const char* szCollection, int fromPos)
{
const char* szFromPos = szCollection + fromPos;
assert(fromPos >= 0);
//assert(szFromPos[0] == ';');
if(pContext->pfnOnNode)
pContext->pfnOnNode(pContext, szFromPos);
if(szFromPos[0] == ';')
{
fromPos++; szFromPos++;
}
while(1)
{
fromPos += skipSpaceChars(szFromPos, NULL);
if(szCollection[fromPos] == '/0' || findchar(";)(", -1, szCollection[fromPos]) >= 0)
break;
fromPos = parseProperty(pContext, szCollection, fromPos);
szFromPos = szCollection + fromPos;
}
return fromPos;
}
第三步，解析节点序列（parseNodeSequence）。节点的顺序排列，至少有一个节点，后面可能还有0个或多个节点。仍然是一个while循环搞定：
//NodeSequence: node{node}
int parseNodeSequence(SGFParseContext* pContext, const char* szCollection, int fromPos)
{
const char* szFromPos = szCollection + fromPos;
assert(fromPos >= 0);
//assert(szFromPos[0] == ';');
while(1)
{
fromPos = parseNode(pContext, szCollection, fromPos);
fromPos += skipSpaceChars(szFromPos, NULL);
szFromPos = szCollection + fromPos;
if(szFromPos[0] != ';')
{
if(pContext->pfnOnNodeEnd)
pContext->pfnOnNodeEnd(pContext);
break;
}
}
return fromPos;
}
第四步，解析树（parseGameTree）。树是一个嵌套结构，最外层是一对括号“(”“)”，里面是N个节点序列或N个嵌套的子树。仍然用一个while循环搞定，遇到“(”则递归调用parseGameTree()解析树或其子树，否则调用parseNodeSequence()解析节点序列。代码如下：
//GameTree: ( {[NodeSequence]|[GameTree]} )
//old GameTree: ( NodeSequence {GameTree} )
int parseGameTree(SGFParseContext* pContext, const char* szCollection, int fromPos)
{
char c;
const char* szFromPos = szCollection + fromPos;
assert(fromPos >= 0);
assert(szFromPos[0] == '(');
pContext->treeIndex++;
if(pContext->pfnOnTree)
pContext->pfnOnTree(pContext, szFromPos, pContext->treeIndex);
fromPos++; szFromPos++;
fromPos += skipSpaceChars(szFromPos, NULL);
c = szCollection[fromPos];
while(1)
{
if(c == '(')
fromPos = parseGameTree(pContext, szCollection, fromPos);
else
fromPos = parseNodeSequence(pContext, szCollection, fromPos);
szFromPos = szCollection + fromPos;
fromPos += skipSpaceChars(szFromPos, NULL);
c = szCollection[fromPos];
if(c == ')')
{
if(pContext->pfnOnTreeEnd)
pContext->pfnOnTreeEnd(pContext, pContext->treeIndex);
pContext->treeIndex--;
break;
}
}
return (fromPos + 1);
}
第五步，最后一步了，解析整个SGF文本内容（parseSGF）。这是对外公开的核心接口。N个树的顺序排列，好办呀，循环调用parseGameTree()顺序解析各个树不就OK了？代码如下：
//SGFCollection: GameTree {GameTree}
int parseSGF(SGFParseContext* pContext, const char* szCollection, int fromPos)
{
const char* szFromPos = szCollection + fromPos;
assert(fromPos >= 0);
assert(szFromPos[0] == '(');
pContext->treeIndex = -1;
while(1)
{
fromPos = parseGameTree(pContext, szCollection, fromPos);
fromPos += skipSpaceChars(szFromPos, NULL);
szFromPos = szCollection + fromPos;
if(szFromPos[0] != '(')
break;
}
return fromPos;
}
测试代码：
int main(int argc, char *argv[])
{
char* s;
int x;
SGFParseContext Context;
//initSGFParseContext(&Context, onTree, onTreeEnd, onNode, onNodeEnd, onProperty, NULL);
initSGFParseContext(&Context, onTree2, onTreeEnd2, onNode2, onNodeEnd2, onProperty2, NULL);
//test parse property:
{
s = "AB[cdef]X[xyz]";
printf("/ntest parse property: ----- /n");
x = parseProperty(&Context, s, 0);
x = parseProperty(&Context, s, 8);
s = "C[ab//]cd]";
x = parseProperty(&Context, s, 0);
}
//test parse node:
{
s = ";A[a]BB[bb]C[]";
printf("/ntest parse node: ----- /n");
x = parseNode(&Context, s, 0);
s = ";A[a];BB[bb]C[]";
x = parseNode(&Context, s, 0);
x = parseNodeSequence(&Context, s, 0);
}
//test parse tree:
{
printf("/ntest parse tree: ----- /n");
s = "(;A[a](;C[c](X[x])Z[z]);D[d](;E[e](F[ff])))";
x = parseGameTree(&Context, s, 0);
}
#if 1
//parse real sgf file:
{
int len = 0;
void* data = NULL;
FILE* pfile = fopen("d://x.txt", "r");
printf("/n---------- test parse real sgf file: -------- /n");
if(pfile)
{
fseek(pfile, 0, SEEK_END);
len = ftell(pfile);
assert(len > 0);
fseek(pfile, 0, SEEK_SET);
data = malloc(len);
assert(data);
fread(data, 1, len, pfile);
parseSGF(&Context, data, 0);
fclose(pfile);
pfile = NULL;
}
}
#endif
{
char c;
printf("/n----- any key to exit: ----- /n");
fflush(stdout);
scanf("%c", &c);
}
}
总结：整个SGF解析器结构比较清晰，只要按照EBNF定义，按部就班地逐步处理即可，不是特别复杂。但由于牵涉到文本、指针、递归，有许多细节需要注意。各位朋友不妨评估一下，自己需要花费多久可以写出类似这样一个SGF解析器？如果时间充裕，也不妨真的动手写一下，看看是否眼高手低呢？所谓的“重复发明轮子”，并非绝对的毫无意义，至少可以锻炼我的动手能力。
另外，有一个设计上的取舍，不知是较好还是较坏。所有的回调函数，目前都有一个 SGFParseContext* pContext ，而此前相同位置的参数是 void* pUserData。是后来考虑到回调函数可能需要访问SGFParseContext中的相关数据（如在PFN_ON_NODE中读取treeIndex），为了方便用户使用才引入pContext参数（用户也可以通过pUserData自行传入pContext，终究是多了一步）。目前的做法，似乎暴露了解析器内部结构（SGFParseContext），又似乎增强了回调函数的稳定性和扩展性（即使不改变函数原形也能通过pContext提供额外参数）。
虽然这个SGF解析器已应用到开源软件“M8围棋谱”(http://code.google.com/p/m8weiqipu/)中，并初步达到了实用目的，但并不能保证该解析器已达到工业强度，其实有不少情况尚未测试到，必然会有疏忽错漏之处，诚请各位朋友批评指正。
另注，考虑到与现有SGF格式文件的兼容性，对SGF规范中的EBNF稍做了一定扩展。
完整源代码请参见：
http://code.google.com/p/m8weiqipu/source/browse/trunk/sgf.h
http://code.google.com/p/m8weiqipu/source/browse/trunk/sgf.c
#ifndef __SGF_H__
#define __SGF_H__
// SGF parser, by liigo
// email: com.liigo_at_gmail.com
// blog: http://blog.csdn.net/liigo
// www.tianqiweiqi.com
#ifdef __cplusplus
extern "C" {
#endif
typedef struct _tagSGFParseContext SGFParseContext;
typedef void (*PFN_ON_PROPERTY) (SGFParseContext* pContext, const char* szID, const char* szValue);
typedef void (*PFN_ON_NODE) (SGFParseContext* pContext, const char* szNodeHeader);
typedef void (*PFN_ON_NODE_END) (SGFParseContext* pContext);
typedef void (*PFN_ON_TREE) (SGFParseContext* pContext, const char* szTreeHeader, int treeIndex);
typedef void (*PFN_ON_TREE_END) (SGFParseContext* pContext, int treeIndex);
typedef struct _tagSGFParseContext
{
void* pUserData;
int treeIndex;
PFN_ON_TREE pfnOnTree;
PFN_ON_TREE_END pfnOnTreeEnd;
PFN_ON_NODE pfnOnNode;
PFN_ON_NODE_END pfnOnNodeEnd;
PFN_ON_PROPERTY pfnOnProperty;
char idBuffer[16];
char* valueBuffer;
int valueBufferSize;
}
SGFParseContext;
void initSGFParseContext(SGFParseContext* pContext,
PFN_ON_TREE pfnOnTree, PFN_ON_TREE_END pfnOnTreeEnd,
PFN_ON_NODE pfnOnNode, PFN_ON_NODE_END pfnOnNodeEnd,
PFN_ON_PROPERTY pfnOnProperty,
void* pUserData);
void cleanupSGFParseContext(SGFParseContext* pContext);
int parseSGF(SGFParseContext* pContext, const char* szCollection, int fromPos);
int parseGameTree(SGFParseContext* pContext, const char* szCollection, int fromPos);
int parseNodeSequence(SGFParseContext* pContext, const char* szCollection, int fromPos);
int parseNode(SGFParseContext* pContext, const char* szCollection, int fromPos);
int parseProperty(SGFParseContext* pContext, const char* szCollection, int fromPos);
#ifdef __cplusplus
}
#endif
#endif //__SGF_H__
#include <stdlib.h>
#include <stdio.h>
#include <memory.h>
#include <ctype.h>
#include <assert.h>
#include "sgf.h"
// SGF parser, by liigo
// email: com.liigo_at_gmail.com
// blog: http://blog.csdn.net/liigo
// www.tianqiweiqi.com
int findchar(const char* sz, size_t n, char c)
{
size_t i;
if(n == -1)
{
char x;
i = 0;
while( x = sz[i] )
{
if(x == c) return i;
i++;
}
return -1;
}
else
{
for(i = 0; i < n; i++)
{
if(sz[i] == c) return i;
}
return -1;
}
}
//return the first non-space char's index
int skipSpaceChars(const char* s, char* endchars)
{
const char* p = s;
char c;
assert(s);
while(c = *p)
{
if(endchars)
{
if(findchar(endchars, -1, c) >= 0)
return (p - s);
}
if(!isspace(c)) return (p - s);
p++;
}
return (p - s);
}
//skipChars or endchars can be NULL, but can not all NULL
int skipChars(const char* s, char* skipChars, char* endchars)
{
const char* p = s;
char c;
assert(s && (skipChars || endchars));
while(c = *p)
{
if(skipChars)
{
if(findchar(skipChars, -1, c) == -1)
return (p - s);
}
if(endchars)
{
if(findchar(endchars, -1, c) >= 0)
return (p - s);
}
p++;
}
return (p - s);
}
int isTextPropertyID(const char* szID)
{
static const char* textIDs[] = { "C", "N", "AP", "CA", "AN", "BR", "BT", "CP", "DT", "EV",
"GN", "GC", "ON", "OT", "PB", "PC", "PW", "RE", "RO", "RU",
"SO", "US", "WR", "WT", "FG", "LB" };
static int n = sizeof(textIDs) / sizeof(textIDs[0]);
int i;
for(i = 0; i < n; i++)
{
if(_stricmp(szID, textIDs[i]) == 0)
return 1;
}
return 0;
}
void getEnoughBuffer(SGFParseContext* pContext, int nNeedBufferSize)
{
if(pContext->valueBuffer == NULL || nNeedBufferSize > pContext->valueBufferSize)
{
int buffSize = pContext->valueBufferSize;
if(buffSize == 0) buffSize = 1024;
while(buffSize < nNeedBufferSize) buffSize *= 2;
if(pContext->valueBuffer == NULL)
pContext->valueBuffer = malloc(buffSize + 4);
else
pContext->valueBuffer = realloc(pContext->valueBuffer, buffSize + 4);
pContext->valueBufferSize = buffSize;
}
}
// [value]
int parsePropertyValue(SGFParseContext* pContext, const char* szCollection, int fromPos)
{
const char* szFromPos = szCollection + fromPos;
assert(szFromPos[0] == '[');
if(!isTextPropertyID(pContext->idBuffer))
{
int rindex = findchar(szFromPos, -1, ']');
int nNeedBufferSize = rindex - 1;
assert(rindex >= 0);
getEnoughBuffer(pContext, nNeedBufferSize);
memcpy(pContext->valueBuffer, szFromPos + 1, nNeedBufferSize);
pContext->valueBuffer[nNeedBufferSize] = '\0';
if(pContext->pfnOnProperty)
pContext->pfnOnProperty(pContext, pContext->idBuffer, pContext->valueBuffer);
return (fromPos + rindex + 1);
}
else
{
//parse the text or simple-text value, consider the '\' escape character
const char* s = szFromPos + 1;
char c;
int in_escape = 0;
int valuelen = 0;
getEnoughBuffer(pContext, 1024);
pContext->valueBuffer[0] = '\0';
while(1)
{
c = *s;
assert(c);
if(!in_escape)
{
if(c == '\\')
{
in_escape = 1;
}
else if(c == ']')
{
break;
}
else
{
getEnoughBuffer(pContext, valuelen + 1);
pContext->valueBuffer[valuelen++] = c;
}
}
else
{
//ignore the newline after '\'
if(c != '\r' && c != '\n')
{
getEnoughBuffer(pContext, valuelen + 1);
pContext->valueBuffer[valuelen++] = c;
}
else
{
char nc = *(s+1);
if(nc)
{
if((c=='\r' && nc=='\n') || (c=='\n' && nc=='\r'))
s++;
}
}
in_escape = 0;
}
s++;
}
getEnoughBuffer(pContext, valuelen + 1);
pContext->valueBuffer[valuelen] = '\0';
if(pContext->pfnOnProperty)
pContext->pfnOnProperty(pContext, pContext->idBuffer, pContext->valueBuffer);
return (s - szCollection + 1);
}
}
//Property: id [value] { [value] }
int parseProperty(SGFParseContext* pContext, const char* szCollection, int fromPos)
{
const char* szFromPos;
int lindex;
int nIDBufferSize = sizeof(pContext->idBuffer) - 1;
assert(szCollection && fromPos >= 0);
szFromPos = szCollection + fromPos;
lindex = findchar(szFromPos, -1, '[');
assert(lindex > 0 && lindex < nIDBufferSize);
if(lindex > 0 && lindex < nIDBufferSize)
{
memcpy(pContext->idBuffer, szFromPos, lindex);
pContext->idBuffer[lindex] = '\0';
fromPos = parsePropertyValue(pContext, szCollection, fromPos + lindex);
szFromPos = szCollection + fromPos;
while(1)
{
fromPos += skipSpaceChars(szFromPos, NULL);
if(szCollection[fromPos] != '[')
break;
fromPos = parsePropertyValue(pContext, szCollection, fromPos);
szFromPos = szCollection + fromPos;
}
return fromPos;
}
return -1;
}
//Node: ; {property}
int parseNode(SGFParseContext* pContext, const char* szCollection, int fromPos)
{
const char* szFromPos = szCollection + fromPos;
assert(fromPos >= 0);
//assert(szFromPos[0] == ';');
if(pContext->pfnOnNode)
pContext->pfnOnNode(pContext, szFromPos);
if(szFromPos[0] == ';')
{
fromPos++; szFromPos++;
}
while(1)
{
fromPos += skipSpaceChars(szFromPos, NULL);
if(szCollection[fromPos] == '\0' || findchar(";)(", -1, szCollection[fromPos]) >= 0)
break;
fromPos = parseProperty(pContext, szCollection, fromPos);
szFromPos = szCollection + fromPos;
}
return fromPos;
}
//NodeSequence: node{node}
int parseNodeSequence(SGFParseContext* pContext, const char* szCollection, int fromPos)
{
const char* szFromPos = szCollection + fromPos;
assert(fromPos >= 0);
//assert(szFromPos[0] == ';');
while(1)
{
fromPos = parseNode(pContext, szCollection, fromPos);
fromPos += skipSpaceChars(szFromPos, NULL);
szFromPos = szCollection + fromPos;
if(szFromPos[0] != ';')
{
if(pContext->pfnOnNodeEnd)
pContext->pfnOnNodeEnd(pContext);
break;
}
}
return fromPos;
}
//GameTree: ( {[NodeSequence]|[GameTree]} )
//old GameTree: ( NodeSequence {GameTree} )
int parseGameTree(SGFParseContext* pContext, const char* szCollection, int fromPos)
{
char c;
const char* szFromPos = szCollection + fromPos;
assert(fromPos >= 0);
assert(szFromPos[0] == '(');
pContext->treeIndex++;
if(pContext->pfnOnTree)
pContext->pfnOnTree(pContext, szFromPos, pContext->treeIndex);
fromPos++; szFromPos++;
fromPos += skipSpaceChars(szFromPos, NULL);
c = szCollection[fromPos];
while(1)
{
if(c == '(')
fromPos = parseGameTree(pContext, szCollection, fromPos);
else
fromPos = parseNodeSequence(pContext, szCollection, fromPos);
szFromPos = szCollection + fromPos;
fromPos += skipSpaceChars(szFromPos, NULL);
c = szCollection[fromPos];
if(c == ')')
{
if(pContext->pfnOnTreeEnd)
pContext->pfnOnTreeEnd(pContext, pContext->treeIndex);
pContext->treeIndex--;
break;
}
}
return (fromPos + 1);
}
//SGFCollection: GameTree {GameTree}
int parseSGF(SGFParseContext* pContext, const char* szCollection, int fromPos)
{
const char* szFromPos = szCollection + fromPos;
assert(fromPos >= 0);
assert(szFromPos[0] == '(');
pContext->treeIndex = -1;
while(1)
{
fromPos = parseGameTree(pContext, szCollection, fromPos);
fromPos += skipSpaceChars(szFromPos, NULL);
szFromPos = szCollection + fromPos;
if(szFromPos[0] != '(')
break;
}
return fromPos;
}
void initSGFParseContext(SGFParseContext* pContext,
PFN_ON_TREE pfnOnTree, PFN_ON_TREE_END pfnOnTreeEnd,
PFN_ON_NODE pfnOnNode, PFN_ON_NODE_END pfnOnNodeEnd,
PFN_ON_PROPERTY pfnOnProperty,
void* pUserData)
{
assert(pContext);
pContext->pfnOnTree = pfnOnTree;
pContext->pfnOnTreeEnd = pfnOnTreeEnd;
pContext->pfnOnNode = pfnOnNode;
pContext->pfnOnNodeEnd = pfnOnNodeEnd;
pContext->pfnOnProperty = pfnOnProperty;
pContext->pUserData = pUserData;
pContext->idBuffer[0] = '\0';
pContext->valueBuffer = NULL;
pContext->valueBufferSize = 0;
pContext->treeIndex = -1;
}
void cleanupSGFParseContext(SGFParseContext* pContext)
{
if(pContext->valueBuffer)
free(pContext->valueBuffer);
memset(pContext, 0, sizeof(SGFParseContext));
}
//
//-----------------------------------------------------------------------------
// 以下是测试代码
static void printheader(const char* s, int n)
{
int i;
for(i = 0; i < n; i++)
{
if(s[i] == '\0') break;
printf("%c", s[i]);
}
printf(" ... \n");
fflush(stdout);
}
static void onProperty(SGFParseContext* pContext, const char* szID, const char* szValue)
{
printf(" onProperty: %s[%s] \n", szID, szValue);
fflush(stdout);
}
static void onNode(SGFParseContext* pContext, const char* szNodeHeader)
{
printf(" onNode: ");
printheader(szNodeHeader, 32);
}
static void onNodeEnd(SGFParseContext* pContext)
{
printf(" onNodeEnd \n");
}
static void onTree(SGFParseContext* pContext, const char* szTreeHeader, int treeIndex)
{
printf(" onTree: ");
printheader(szTreeHeader, 32);
}
static void onTreeEnd(SGFParseContext* pContext, int treeIndex)
{
printf(" onTreeEnd \n");
}
//------------------
static void onProperty2(SGFParseContext* pContext, const char* szID, const char* szValue)
{
printf("%s[%s] ", szID, szValue);
//fflush(stdout);
}
static void onNode2(SGFParseContext* pContext, const char* szNodeHeader)
{
printf("; ");
}
static void onNodeEnd2(SGFParseContext* pContext)
{
}
static void onTree2(SGFParseContext* pContext, const char* szTreeHeader, int treeIndex)
{
int i;
printf("\n");
for(i = 0; i < treeIndex; i++)
printf(" ");
printf("[tree:%d]( ", treeIndex);
}
static void onTreeEnd2(SGFParseContext* pContext, int treeIndex)
{
printf(") [end of tree %d]\n\n", treeIndex);
}
static void testParseSGFFile(SGFParseContext* pContext, const char* szFileName)
{
int len = 0;
void* data = NULL;
FILE* pfile = fopen(szFileName, "r");
printf("\n\n--- test parse sgf file: %s ---\n", szFileName);
if(pfile)
{
fseek(pfile, 0, SEEK_END);
len = ftell(pfile);
assert(len > 0);
fseek(pfile, 0, SEEK_SET);
data = malloc(len);
assert(data);
fread(data, 1, len, pfile);
parseSGF(pContext, data, 0);
fclose(pfile);
pfile = NULL;
}
else
{
printf("\n\n--- open sgf file error: %s ---\n\n", szFileName);
}
}
////////////////////
int main(int argc, char *argv[])
{
char* s;
int x;
SGFParseContext Context;
//initSGFParseContext(&Context, onTree, onTreeEnd, onNode, onNodeEnd, onProperty, NULL);
initSGFParseContext(&Context, onTree2, onTreeEnd2, onNode2, onNodeEnd2, onProperty2, NULL);
//test parse property:
{
s = "A[a]B[b1][b2]X[xyz]";
printf("\ntest parse property: ----- \n");
x = parseProperty(&Context, s, 0);
x = parseProperty(&Context, s, 4);
s = "C[ab\\]cd]";
x = parseProperty(&Context, s, 0);
}
//test parse node:
{
s = ";A[a]BB[bb]C[]";
printf("\ntest parse node: ----- \n");
x = parseNode(&Context, s, 0);
s = ";A[a];BB[bb]C[]";
x = parseNode(&Context, s, 0);
x = parseNodeSequence(&Context, s, 0);
}
//test parse tree:
{
printf("\ntest parse tree: ----- \n");
s = "(;A[a](;C[c](X[x])Z[z]);D[d](;E[e](F[ff])))";
x = parseGameTree(&Context, s, 0);
}
#if 1
//parse real sgf file:
{
char filename[256];
int i;
for(i = 1; i <= 7; i++)
{
sprintf(filename, "%d.sgf", i);
testParseSGFFile(&Context, filename);
}
for(i = 0; i <= 365; i++)
{
sprintf(filename, "WuQingYuan\\wqy (%d).sgf", i);
testParseSGFFile(&Context, filename);
}
for(i = 1; i <= 465; i++)
{
sprintf(filename, "标准中国流645局\\S%05d.sgf", i);
testParseSGFFile(&Context, filename);
}
}
#endif
{
char c;
printf("\n----- any key to exit: ----- \n");
fflush(stdout);
scanf("%c", &c);
}
}
