diff src/kaigo/superFC/fcgo.go @ 65:0369656be06c default tip

many changes.
author pyon@macmini
date Fri, 20 May 2022 06:30:34 +0900
parents
children
line wrap: on
line diff
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/src/kaigo/superFC/fcgo.go	Fri May 20 06:30:34 2022 +0900
@@ -0,0 +1,484 @@
+/*
+ fcgo.go: Fucho de Go
+
+ Last Change: 2021-06-14 月 16:14:04.
+*/
+package main
+
+/*
+#cgo LDFLAGS: -L. -lxdwapi -static
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <io.h>
+#include <windows.h>
+#include <xdw_api.h>
+#include <xdwapian.h>
+
+#define MAXLINE 12000
+#define BLOCKSZ   128
+
+void xdw2txt(const char* xdwfile, const char* txtfile) {
+	char in_path[_MAX_PATH];
+    _fullpath(in_path, xdwfile, _MAX_PATH);
+
+	XDW_DOCUMENT_HANDLE h = NULL;
+	XDW_OPEN_MODE_EX mode = {sizeof(XDW_OPEN_MODE_EX), XDW_OPEN_READONLY, XDW_AUTH_NODIALOGUE};
+	if (XDW_OpenDocumentHandle(in_path, &h, (XDW_OPEN_MODE*)&mode)) {
+		printf("XDW Error: cannot open %s\n", xdwfile);
+		return;
+	}
+
+	int api_result = XDW_GetFullText(h, txtfile, NULL); 
+	if (api_result < 0) {
+		printf("Error: cannot write text\n");
+		return;
+	}
+	XDW_CloseDocumentHandle(h, NULL);
+}
+
+void xdwsplit1(const char* xdwfile) {
+	char in_path[_MAX_PATH];
+    _fullpath(in_path, xdwfile, _MAX_PATH);
+
+	XDW_DOCUMENT_HANDLE h = NULL;
+	XDW_OPEN_MODE_EX mode = {sizeof(XDW_OPEN_MODE_EX), XDW_OPEN_READONLY, XDW_AUTH_NODIALOGUE};
+	if (XDW_OpenDocumentHandle(in_path, &h, (XDW_OPEN_MODE*)&mode)) {
+		printf("Error: cannot open %s\n", xdwfile);
+		return;
+	}
+
+	XDW_DOCUMENT_INFO info = {sizeof(XDW_DOCUMENT_INFO), 0, 0, 0};
+	XDW_GetDocumentInformation(h, &info);
+    int nPage = info.nPages;
+
+	char buf[_MAX_PATH];
+    for (int i = 1; i <= nPage; i++) {
+		sprintf(buf, "fctmp_%05d.xdw", i);
+		_fullpath(in_path, buf, _MAX_PATH);
+
+		int api_result = XDW_GetPage(h, i, in_path, NULL);
+		if (api_result < 0) {
+			printf("XDW Error: cannot get page\n");
+			return;
+		}
+	}
+	XDW_CloseDocumentHandle(h, NULL);
+}
+
+void xdwmerge(const char* list, const char* output) {
+	
+	FILE *fp;
+	if ((fp = fopen(list, "r")) == NULL) {
+		fprintf(stderr, "can't open file [%s]\n", list);
+		exit(1);
+	}
+
+	char *all_path = (char*)malloc(MAXLINE * sizeof(char) * _MAX_PATH);
+	if (all_path == NULL) {
+		fprintf(stderr, "can't allocate memory\n");
+		exit(1);
+	}
+
+	int n = 0;
+	char *q;
+	char buf[_MAX_PATH];
+
+	while (fgets(buf, sizeof buf, fp)) {
+		if ((q = strchr(buf, '\n')) != NULL) {
+			*q = '\0';
+		}
+		_fullpath(buf, buf, _MAX_PATH);
+		strncpy(&all_path[n * _MAX_PATH], buf, _MAX_PATH);
+		n++;
+	}
+	fclose(fp);
+
+	char *blk_path = (char*)malloc(BLOCKSZ * sizeof(char) * _MAX_PATH);
+	const char **blk_path_addr = (const char**)malloc((n / BLOCKSZ + 1) * sizeof(char*) * _MAX_PATH);
+	if (blk_path == NULL || blk_path_addr == NULL) {
+		fprintf(stderr, "can't allocate memory\n");
+		exit(1);
+	}
+
+	// process by block
+	int api_result;
+	int bn = 0;
+	for (int p = 0, m = 0; p < n; p++) {
+		m = p % BLOCKSZ;
+		if (m == 0 && p > 0) {
+			sprintf(buf, "fctmp_b%04d.xdw", ++bn);
+			_fullpath(buf, buf, _MAX_PATH);
+			api_result = XDW_MergeXdwFiles(blk_path_addr, BLOCKSZ, buf, NULL);
+			if (api_result < 0) {
+				fprintf(stderr, "can't merge [1] (p = %d, m = %d)\n", p, m);
+				exit(1);
+			}
+		} 
+		strncpy(&blk_path[m * _MAX_PATH], &all_path[p * _MAX_PATH], _MAX_PATH);
+		blk_path_addr[m] = &blk_path[m * _MAX_PATH];
+	}
+
+	sprintf(buf, "fctmp_b%04d.xdw", ++bn);
+	_fullpath(buf, buf, _MAX_PATH);
+	int mod = n % BLOCKSZ;
+	if (mod == 0) mod = BLOCKSZ; 
+	api_result = XDW_MergeXdwFiles(blk_path_addr, mod, buf, NULL);
+	if (api_result < 0) {
+		fprintf(stderr, "can't merge [2]\n");
+		exit(1);
+	}
+
+	// merge blocks
+	for (int b = 0; b < bn; b++) {
+		sprintf(buf, "fctmp_b%04d.xdw", b + 1);
+		_fullpath(buf, buf, _MAX_PATH);
+		strncpy(&blk_path[b * _MAX_PATH], buf, _MAX_PATH);
+		blk_path_addr[b] = &blk_path[b * _MAX_PATH];
+	}
+	_fullpath(buf, output, _MAX_PATH );
+	api_result = XDW_MergeXdwFiles(blk_path_addr, bn, buf, NULL);
+	if (api_result < 0) {
+		fprintf(stderr, "can't merge [3]\n");
+		exit(1);
+	}
+
+	free(all_path);
+	free(blk_path);
+	free(blk_path_addr);
+}
+
+int xdwaddannotation(XDW_DOCUMENT_HANDLE h, int page, int x, int y, char* string, int* sz, int tr)
+{
+	XDW_ANNOTATION_HANDLE annoation;
+	int api_result = XDW_AddAnnotation(h, XDW_AID_TEXT, page, x, y, NULL, &annoation, NULL);
+	if (api_result < 0) {
+		fprintf(stderr, "can't make annotation\n");
+		return -1;
+	}
+
+	api_result = XDW_SetAnnotationAttribute(h, annoation, XDW_ATN_Text, XDW_ATYPE_STRING, string, 0, NULL);
+	api_result = XDW_SetAnnotationAttribute(h, annoation, XDW_ATN_FontSize, XDW_ATYPE_INT, (char*)(sz), 0, NULL);
+    if (tr) {
+		int color = XDW_COLOR_NONE;
+		api_result = XDW_SetAnnotationAttribute(h, annoation, XDW_ATN_BackColor, XDW_ATYPE_INT, (char*)(&color), 0, NULL);
+    }
+	return 0;
+}
+
+void xdwaddpage(const char* file, int sp) {
+	XDW_DOCUMENT_HANDLE h = NULL;
+	XDW_OPEN_MODE_EX mode = {sizeof(XDW_OPEN_MODE_EX), XDW_OPEN_UPDATE, XDW_AUTH_NODIALOGUE};
+	int api_result = XDW_OpenDocumentHandle(file, &h, (XDW_OPEN_MODE*)&mode);
+	if (api_result < 0) {
+		fprintf(stderr, "can't open file\n");
+		exit(1);
+	}
+
+	XDW_DOCUMENT_INFO info = {sizeof(XDW_DOCUMENT_INFO), 0};
+	XDW_GetDocumentInformation(h, &info);
+	int last_page = info.nPages;
+
+    int sz = 80;
+    int tr = 1;
+    char pagenum[12];
+
+    for (int p = 0; p < last_page; p++) {
+        sprintf(pagenum, "%05d-%05d", p + sp, last_page);
+        api_result = xdwaddannotation(h, p + 1, 1769, 5658, pagenum, &sz, tr);
+		api_result = xdwaddannotation(h, p + 1, 18297, 28326, pagenum, &sz, tr);
+        if (api_result < 0) break;
+    }
+
+	if (api_result >= 0) api_result = XDW_SaveDocument(h, NULL);
+
+	XDW_CloseDocumentHandle(h, NULL);
+}
+
+
+*/
+import "C"
+
+import (
+	"encoding/csv"
+	"flag"
+	"fmt"
+	"io"
+	"log"
+	"os"
+	"regexp"
+	"strings"
+	"time"
+
+	"golang.org/x/text/encoding/japanese"
+	"golang.org/x/text/transform"
+)
+
+type Data struct {
+	Hno string
+	Page string
+}
+
+func (d *Data) ToCsv() string {
+	s := []string{d.Hno, d.Page}
+	return strings.Join(s, ",")
+}
+
+var (
+    ver = "0.1"
+
+	clean bool
+	skip bool
+	debug bool
+	xdwfile string
+
+	txtfile = "fctmp_txt.txt"
+	infofile = "fctmp_info.csv"
+	orderfile = "sort.list"	// input
+	pagefile = "fctmp_page.txt"
+	output = "output.xdw"
+)
+
+func init() {
+	/* INITIALIZE FLAGS */
+	flag.BoolVar(&clean, "c", false, "clean temporary files & exit")
+	flag.BoolVar(&skip, "e", false, "use existed files")
+	flag.BoolVar(&debug, "d", false, "debug mode")
+	flag.StringVar(&xdwfile, "i", "fc.xdw", "target xdw file")
+}
+
+func main() {
+	flag.Parse()
+
+    /* PRINT HEADER */
+    fmt.Println("===================================================")
+    fmt.Println(" 普徴のみならず納通なら被保番でソートできるかも... ")
+    fmt.Printf("         - fcgo [ver %s] -\n", ver)
+    fmt.Println("===================================================\n")
+	print_time("now")
+	fmt.Println("[0] start ...")
+
+    /* CLEAN TEMPORARY DIRECTORY */
+	os.Remove(output)
+	if !skip {
+		clean_full()
+	}
+	if clean {
+		os.Exit(0)
+	}
+	print_time("check done")
+
+	fmt.Println("[1] extract ...")
+	if !skip {
+		C.xdw2txt(C.CString(xdwfile), C.CString(txtfile))
+	}
+	print_time("extract done.")
+
+	c := make(chan int)
+	fmt.Println("[2] split ... (run background)")
+	go func() {
+		if !skip {
+			C.xdwsplit1(C.CString(xdwfile))
+		}
+		print_time("split done.")
+		c <- 1
+	}()
+
+	fmt.Println("[3] analize ...")
+	data_hash, err := analize(txtfile, infofile)
+	if err != nil {
+		log.Fatal(err)
+	}
+	print_time("analize done.")
+	debug_print(debug, fmt.Sprintf("len = %d", len(data_hash)))
+
+	fmt.Println("[4] read order ...")
+	h_order, n, err := read_order(orderfile)
+	if err != nil {
+		log.Fatal(err)
+	}
+	print_time("read order done.")
+	debug_print(debug, fmt.Sprintf("len = %d / %d", len(h_order), n))
+
+	fmt.Println("[5] make_list ...")
+	n, err = make_list(h_order, data_hash, pagefile)
+	if err != nil {
+		log.Fatal(err)
+	}
+	print_time("make list done.")
+	debug_print(debug, fmt.Sprintf("order = %d", n))
+
+	<-c
+
+	fmt.Println("[6] merge ...")
+	clean_mini()
+	C.xdwmerge(C.CString(pagefile), C.CString(output))
+	print_time("merge done.")
+	clean_mini()
+
+	fmt.Println("[7] page ...")
+	C.xdwaddpage(C.CString(output), C.int(1))
+	print_time("page done.")
+
+	if !debug {
+		fmt.Println("[8] clean ...")
+		clean_full()
+		print_time("clean done.")
+	}
+}
+
+func analize(txtfile, infofile string) (map[string]Data, error) {
+	hash := make(map[string]Data)
+
+	c, err := os.ReadFile(txtfile)
+	if err != nil {
+		return hash, err
+	}
+	r := strings.NewReader(string(c))
+	tr := transform.NewReader(r, japanese.ShiftJIS.NewDecoder())
+	b, err := io.ReadAll(tr)
+	if err != nil {
+		return hash, err
+	}
+
+	regHno := regexp.MustCompile(`0[1238]0[0-9]{7}`)
+
+	var csv string
+
+	buf := strings.Split(string(b), "期別保険料額")
+	for p, v := range buf[:len(buf)-1] {
+		s := zen2han(v)
+		hno := regHno.FindString(s)
+		page := fmt.Sprintf("%05d", p + 1)
+
+		data := Data {
+			Hno: hno,
+			Page: page,
+		}
+		hash[hno] = data
+
+		csv += data.ToCsv() + "\n"
+	}
+
+	if err := os.WriteFile(infofile, []byte(csv), 0644); err != nil {
+		return hash, err
+	}
+	return hash, nil;
+}
+
+func zen2han(s string) (string) {
+	s = strings.ReplaceAll(s, "0", "0")
+	s = strings.ReplaceAll(s, "1", "1")
+	s = strings.ReplaceAll(s, "2", "2")
+	s = strings.ReplaceAll(s, "3", "3")
+	s = strings.ReplaceAll(s, "4", "4")
+	s = strings.ReplaceAll(s, "5", "5")
+	s = strings.ReplaceAll(s, "6", "6")
+	s = strings.ReplaceAll(s, "7", "7")
+	s = strings.ReplaceAll(s, "8", "8")
+	return strings.ReplaceAll(s, "9", "9")
+}
+
+func read_order(csvfile string) ([]string, int, error) {
+	var order []string
+
+	c, err := os.ReadFile(csvfile)
+	if err != nil {
+		return order, -1, err
+	}
+	r := strings.NewReader(string(c))
+	tr := transform.NewReader(r, japanese.ShiftJIS.NewDecoder())
+	b, err := io.ReadAll(tr)
+	if err != nil {
+		return order, -1, err
+	}
+
+	n := 0
+	cr := csv.NewReader(strings.NewReader(string(b)))
+	for {
+		record, err := cr.Read()
+		if err == io.EOF {
+			break
+		}
+		if err != nil {
+			return order, n, err
+		}
+
+		s := zen2han(record[0])
+		if strings.HasPrefix(s, "0") {
+			order = append(order, s)
+		}
+		n++
+	}
+	return order, n, nil
+}
+
+func make_list(h_order []string, hash map[string]Data, pagefile string) (int, error) {
+	var n int
+	var list []string
+	done := make(map[string]bool)
+
+	for _, h := range h_order {
+		if _, ok := done[h]; !ok {
+			if data, ok := hash[h]; ok {
+				list = append(list, data.Page)
+				done[h] = true
+				n++
+			}
+		}
+	}
+
+	if err := write_pagefile(pagefile, list); err != nil {
+		return n, err
+	}
+	return n, nil
+}
+
+func write_pagefile(file string, list []string) error {
+	f, err := os.OpenFile(file, os.O_RDWR|os.O_CREATE, 0755)
+	if err != nil {
+		return err
+	}
+	for _, p := range list {
+		fmt.Fprintf(f, "fctmp_%s.xdw\n", p)
+	}
+	if err := f.Close(); err != nil {
+		return err
+	}
+	return nil
+}
+
+func print_time(msg string) {
+	now := time.Now()
+	fmt.Printf("\t%v # %s\n", now, msg)
+}
+
+func debug_print(debug bool, msg string) {
+	if debug {
+		fmt.Printf("\t%s\n", msg)
+	}
+}
+
+func clean_full() error {
+	return clean_file("fctmp_")
+}
+
+func clean_mini() error {
+	return clean_file("fctmp_b")
+}
+
+func clean_file(prefix string) error {
+	files, err := os.ReadDir(".")
+	if err != nil {
+		return err
+	}
+
+	for _, file := range files {
+		if strings.HasPrefix(file.Name(), prefix) {
+			os.Remove(file.Name())
+			continue
+		}
+	}
+	return err
+}
+