X-Git-Url: http://pileus.org/git/?p=grits;a=blobdiff_plain;f=src%2Fdata%2Fgis-http.c;h=b5259e47108e15a92f0f95b44c24d4b19828a873;hp=78251bec02b8230a66f6719f2ba19ffc6d49e4db;hb=e7127ea74717754c48779467aca2f9a1596a99f0;hpb=7d7d6ae0531a85361ce0d8b997d988b653282847
diff --git a/src/data/gis-http.c b/src/data/gis-http.c
index 78251be..b5259e4 100644
--- a/src/data/gis-http.c
+++ b/src/data/gis-http.c
@@ -15,6 +15,16 @@
* along with this program. If not, see .
*/
+/**
+ * SECTION:gis-http
+ * @short_description: Hyper Text Transfer Protocol
+ *
+ * #GisHttp is a small wrapper around libsoup to provide data access using the
+ * Hyper Text Transfer Protocol. Each #GisHttp should be associated with a
+ * particular server or dataset, all the files downloaded for this dataset will
+ * be cached together in $HOME.cache/libgis/
+ */
+
#include
#include
#include
@@ -22,6 +32,21 @@
#include "gis-http.h"
+gchar *_get_cache_path(GisHttp *http, const gchar *local)
+{
+ return g_build_filename(g_get_user_cache_dir(), PACKAGE,
+ http->prefix, local, NULL);
+}
+
+/**
+ * gis_http_new:
+ * @prefix: The prefix in the cache to store the downloaded files.
+ * For example: * "/nexrad/level2/".
+ *
+ * Create a new #GisHttp for the given prefix
+ *
+ * Returns: the new #GisHttp
+ */
GisHttp *gis_http_new(const gchar *prefix)
{
g_debug("GisHttp: new - %s", prefix);
@@ -32,6 +57,12 @@ GisHttp *gis_http_new(const gchar *prefix)
return http;
}
+/**
+ * gis_http_free:
+ * @http: the #GisHttp to free
+ *
+ * Frees resources used by @http and cancels any pending requests.
+ */
void gis_http_free(GisHttp *http)
{
g_debug("GisHttp: free - %s", http->prefix);
@@ -42,7 +73,7 @@ void gis_http_free(GisHttp *http)
}
/* For passing data to the chunck callback */
-struct _cache_info {
+struct _CacheInfo {
FILE *fp;
gchar *path;
GisChunkCallback callback;
@@ -54,7 +85,7 @@ struct _cache_info {
*/
static void _chunk_cb(SoupMessage *message, SoupBuffer *chunk, gpointer _info)
{
- struct _cache_info *info = _info;
+ struct _CacheInfo *info = _info;
if (!SOUP_STATUS_IS_SUCCESSFUL(message->status_code)) {
g_warning("GisHttp: _chunk_cb - soup failed with %d",
@@ -74,15 +105,27 @@ static void _chunk_cb(SoupMessage *message, SoupBuffer *chunk, gpointer _info)
}
}
+/**
+ * gis_http_fetch:
+ * @http: the #GisHttp connection to use
+ * @uri: the URI to fetch
+ * @local: the local name to give to the file
+ * @mode: the update type to use when fetching data
+ * @callback: callback to call when a chunk of data is received
+ * @user_data: user data to pass to the callback
+ *
+ * Fetch a file from the cache. Whether the file is actually loaded from the
+ * remote server depends on the value of @mode.
+ *
+ * Returns: The local path to the complete file
+ */
/* TODO: use .part extentions and continue even when using GIS_ONCE */
gchar *gis_http_fetch(GisHttp *http, const gchar *uri, const char *local,
GisCacheType mode, GisChunkCallback callback, gpointer user_data)
{
- g_debug("GisHttp: fetch - %.20s... >> %s/%s mode=%d",
+ g_debug("GisHttp: fetch - %s... >> %s/%s mode=%d",
uri, http->prefix, local, mode);
-
- gchar *path = g_build_filename(g_get_user_cache_dir(), PACKAGE,
- http->prefix, local, NULL);
+ gchar *path = _get_cache_path(http, local);
/* Unlink the file if we're refreshing it */
if (mode == GIS_REFRESH)
@@ -91,10 +134,14 @@ gchar *gis_http_fetch(GisHttp *http, const gchar *uri, const char *local,
/* Do the cache if necessasairy */
if (!(mode == GIS_ONCE && g_file_test(path, G_FILE_TEST_EXISTS)) &&
mode != GIS_LOCAL) {
- g_debug("GisHttp: do_cache - Caching file %s", local);
+ g_debug("GisHttp: fetch - Caching file %s", local);
/* Open the file for writting */
- FILE *fp = fopen_p(path, "a");
+ gchar *part = path;
+ if (!g_file_test(path, G_FILE_TEST_EXISTS))
+ part = g_strdup_printf("%s.part", path);
+ FILE *fp = fopen_p(part, "ab");
+ fseek(fp, 0, SEEK_END); // "a" is broken on Windows, twice
/* Make temp data */
struct _CacheInfo info = {
@@ -112,16 +159,95 @@ gchar *gis_http_fetch(GisHttp *http, const gchar *uri, const char *local,
soup_message_headers_set_range(message->request_headers, ftell(fp), -1);
soup_session_send_message(http->soup, message);
+ /* Close file */
+ fclose(fp);
+ if (path != part && SOUP_STATUS_IS_SUCCESSFUL(message->status_code)) {
+ g_rename(part, path);
+ g_free(part);
+ }
+
/* Finished */
if (message->status_code == 416) {
/* Range unsatisfiable, file already complete */
- } else if (!SOUP_STATUS_IS_SUCCESSFUL(message->status_code))
+ } else if (!SOUP_STATUS_IS_SUCCESSFUL(message->status_code)) {
g_warning("GisHttp: done_cb - error copying file, status=%d\n"
"\tsrc=%s\n"
"\tdst=%s",
message->status_code, uri, path);
+ return NULL;
+ }
}
+
/* TODO: free everything.. */
return path;
}
+
+/**
+ * gis_http_available:
+ * @http: the #GisHttp connection to use
+ * @filter: filter used to extract files from the index, or NULL
+ * For example: "href=\"([^"]*)\""
+ * @cache: path to the local cache, or NULL to not search the cache
+ * @extract: regex used to extract filenames from the page, should match the
+ * filename as $1, or NULL to use /http="([^"])"/
+ * @index: path to the index page, or NULL to not search online
+ *
+ * Look through the cache and an HTTP index page for a list of available files.
+ * The name of each file that matches the filter is added to the returned list.
+ *
+ * The list as well as the strings contained in it should be freed afterwards.
+ *
+ * Returns the list of matching filenames
+ */
+GList *gis_http_available(GisHttp *http,
+ gchar *filter, gchar *cache,
+ gchar *extract, gchar *index)
+{
+ g_debug("GisHttp: available - %s~=%s %s~=%s",
+ filter, cache, extract, index);
+ GRegex *filter_re = g_regex_new(filter, 0, 0, NULL);
+ GList *files = NULL;
+
+ /* Add cached files */
+ if (cache) {
+ const gchar *file;
+ gchar *path = _get_cache_path(http, cache);
+ GDir *dir = g_dir_open(path, 0, NULL);
+ while ((file = g_dir_read_name(dir)))
+ if (g_regex_match(filter_re, file, 0, NULL))
+ files = g_list_prepend(files, g_strdup(file));
+ g_free(path);
+ }
+
+ /* Add online files if online */
+ if (index) {
+ gchar tmp[16];
+ g_snprintf(tmp, sizeof(tmp), ".index.%x", g_random_int());
+ gchar *path = gis_http_fetch(http, index, tmp,
+ GIS_REFRESH, NULL, NULL);
+ gchar *html;
+ g_file_get_contents(path, &html, NULL, NULL);
+
+ /* Match hrefs by default, this regex is not very accurate */
+ GRegex *extract_re = g_regex_new(
+ extract ?: "href=\"([^\"]*)\"", 0, 0, NULL);
+ GMatchInfo *info;
+ g_regex_match(extract_re, html, 0, &info);
+ while (g_match_info_matches(info)) {
+ gchar *file = g_match_info_fetch(info, 1);
+ if (g_regex_match(filter_re, file, 0, NULL))
+ files = g_list_prepend(files, file);
+ else
+ g_free(file);
+ g_match_info_next(info, NULL);
+ }
+
+ g_match_info_free(info);
+ g_unlink(path);
+ g_free(path);
+ g_free(html);
+ }
+
+ return files;
+}