Optionally ES schema from file #117

docs/USAGE.md

@@ -53,6 +53,8 @@ Index options
     --es-index=<str>              Elasticsearch index name. DEFAULT=sist2
     -p, --print                   Just print JSON documents to stdout.
     --script-file=<str>           Path to user script.
+    --mappings-file=<str>         Path to Elasticsearch mappings.
+    --settings-file=<str>         Path to Elasticsearch settings.
     --async-script                Execute user script asynchronously.
     --batch-size=<int>            Index batch size. DEFAULT: 100
     -f, --force-reset             Reset Elasticsearch mappings and settings. (You must use this option the first time you use the index command)
@@ -257,6 +259,10 @@ it is currently unsupported and has no guaranties of back/forward compatibility.
     Print index in JSON format to stdout.
  * `--script-file` 
     Path to user script. See [Scripting](scripting.md).
+ * `--mappings-file`
+    Path to custom Elasticsearch mappings. If none is specified, [the bundled mappings](https://github.com/simon987/sist2/tree/master/schema) will be used.
+ * `--settings-file`
+    Path to custom Elasticsearch settings. *(See above)*
  * `--async-script` 
     Use `wait_for_completion=false` elasticsearch option while executing user script.
      (See [Elasticsearch documentation](https://www.elastic.co/guide/en/elasticsearch/reference/current/tasks.html))
@@ -266,7 +272,6 @@ it is currently unsupported and has no guaranties of back/forward compatibility.
     down the process.
  * `-f, --force-reset` 
     Reset Elasticsearch mappings and settings.
-    **(You must use this option the first time you use the index command)**.
     
 ### Index examples
 

src/cli.c

@@ -56,6 +56,12 @@ void scan_args_destroy(scan_args_t *args) {
 
 void index_args_destroy(index_args_t *args) {
     //todo
+    if (args->es_mappings_path) {
+        free(args->es_mappings);
+    }
+    if (args->es_settings_path) {
+        free(args->es_settings);
+    }
     free(args);
 }
 
@@ -231,25 +237,25 @@ int scan_args_validate(scan_args_t *args, int argc, const char **argv) {
     return 0;
 }
 
-int load_script(const char *script_path, char **dst) {
+int load_external_file(const char *file_path, char **dst) {
     struct stat info;
-    int res = stat(script_path, &info);
+    int res = stat(file_path, &info);
 
     if (res == -1) {
-        fprintf(stderr, "Error opening script file '%s': %s\n", script_path, strerror(errno));
+        LOG_ERRORF("cli.c", "Error opening file '%s': %s\n", file_path, strerror(errno))
         return 1;
     }
 
-    int fd = open(script_path, O_RDONLY);
+    int fd = open(file_path, O_RDONLY);
     if (fd == -1) {
-        fprintf(stderr, "Error opening script file '%s': %s\n", script_path, strerror(errno));
+        LOG_ERRORF("cli.c", "Error opening file '%s': %s\n", file_path, strerror(errno))
         return 1;
     }
 
     *dst = malloc(info.st_size + 1);
     res = read(fd, *dst, info.st_size);
     if (res < 0) {
-        fprintf(stderr, "Error reading script file '%s': %s\n", script_path, strerror(errno));
+        LOG_ERRORF("cli.c", "Error reading file '%s': %s\n", file_path, strerror(errno))
         return 1;
     }
 
@@ -293,7 +299,19 @@ int index_args_validate(index_args_t *args, int argc, const char **argv) {
     }
 
     if (args->script_path != NULL) {
-        if (load_script(args->script_path, &args->script) != 0) {
+        if (load_external_file(args->script_path, &args->script) != 0) {
+            return 1;
+        }
+    }
+
+    if (args->es_settings_path != NULL) {
+        if (load_external_file(args->es_settings_path, &args->es_settings) != 0) {
+            return 1;
+        }
+    }
+
+    if (args->es_mappings_path != NULL) {
+        if (load_external_file(args->es_mappings_path, &args->es_mappings) != 0) {
             return 1;
         }
     }
@@ -309,6 +327,10 @@ int index_args_validate(index_args_t *args, int argc, const char **argv) {
     LOG_DEBUGF("cli.c", "arg async_script=%s", args->async_script)
     LOG_DEBUGF("cli.c", "arg script=%s", args->script)
     LOG_DEBUGF("cli.c", "arg print=%d", args->print)
+    LOG_DEBUGF("cli.c", "arg es_mappings_path=%s", args->es_mappings_path)
+    LOG_DEBUGF("cli.c", "arg es_mappings=%s", args->es_mappings)
+    LOG_DEBUGF("cli.c", "arg es_settings_path=%s", args->es_settings_path)
+    LOG_DEBUGF("cli.c", "arg es_settings=%s", args->es_settings)
     LOG_DEBUGF("cli.c", "arg batch_size=%d", args->batch_size)
     LOG_DEBUGF("cli.c", "arg force_reset=%d", args->force_reset)
 
@@ -445,7 +467,7 @@ int exec_args_validate(exec_args_t *args, int argc, const char **argv) {
         LOG_FATAL("cli.c", "--script-file argument is required");
     }
 
-    if (load_script(args->script_path, &args->script) != 0) {
+    if (load_external_file(args->script_path, &args->script) != 0) {
         return 1;
     }
 

src/cli.h

@@ -39,6 +39,10 @@ typedef struct index_args {
     const char *index_path;
     const char *script_path;
     char *script;
+    const char *es_settings_path;
+    char *es_settings;
+    const char *es_mappings_path;
+    char *es_mappings;
     int print;
     int batch_size;
     int async_script;

src/index/elastic.c

@@ -356,7 +356,7 @@ void finish_indexer(char *script, int async_script, char *index_id) {
     free_response(r);
 }
 
-void elastic_init(int force_reset) {
+void elastic_init(int force_reset, const char* user_mappings, const char* user_settings) {
 
     // Check if index exists
     char url[4096];
@@ -392,13 +392,13 @@ void elastic_init(int force_reset) {
         free_response(r);
 
         snprintf(url, sizeof(url), "%s/%s/_settings", IndexCtx.es_url, IndexCtx.es_index);
-        r = web_put(url, settings_json);
-        LOG_INFOF("elastic.c", "Update settings <%d>", r->status_code);
+        r = web_put(url, user_settings ? user_settings : settings_json);
+        LOG_INFOF("elastic.c", "Update user_settings <%d>", r->status_code);
         free_response(r);
 
         snprintf(url, sizeof(url), "%s/%s/_mappings/_doc?include_type_name=true", IndexCtx.es_url, IndexCtx.es_index);
-        r = web_put(url, mappings_json);
-        LOG_INFOF("elastic.c", "Update mappings <%d>", r->status_code);
+        r = web_put(url, user_mappings ? user_mappings : mappings_json);
+        LOG_INFOF("elastic.c", "Update user_mappings <%d>", r->status_code);
         free_response(r);
 
         snprintf(url, sizeof(url), "%s/%s/_open", IndexCtx.es_url, IndexCtx.es_index);

src/index/elastic.h

@@ -25,7 +25,7 @@ es_indexer_t *create_indexer(const char *url, const char *index);
 void elastic_cleanup();
 void finish_indexer(char *script, int async_script, char *index_id);
 
-void elastic_init(int force_reset);
+void elastic_init(int force_reset, const char* user_mappings, const char* user_settings);
 
 cJSON *elastic_get_document(const char *uuid_str);
 

src/main.c

@@ -267,7 +267,7 @@ void sist2_index(index_args_t *args) {
     IndexCtx.batch_size = args->batch_size;
 
     if (!args->print) {
-        elastic_init(args->force_reset);
+        elastic_init(args->force_reset, args->es_mappings, args->es_settings);
     }
 
     char descriptor_path[PATH_MAX];
@@ -446,6 +446,8 @@ int main(int argc, const char *argv[]) {
             OPT_STRING(0, "es-index", &common_es_index, "Elasticsearch index name. DEFAULT=sist2"),
             OPT_BOOLEAN('p', "print", &index_args->print, "Just print JSON documents to stdout."),
             OPT_STRING(0, "script-file", &common_script_path, "Path to user script."),
+            OPT_STRING(0, "mappings-file", &index_args->es_mappings_path, "Path to Elasticsearch mappings."),
+            OPT_STRING(0, "settings-file", &index_args->es_settings_path, "Path to Elasticsearch settings."),
             OPT_BOOLEAN(0, "async-script", &common_async_script, "Execute user script asynchronously."),
             OPT_INTEGER(0, "batch-size", &index_args->batch_size, "Index batch size. DEFAULT: 100"),
             OPT_BOOLEAN('f', "force-reset", &index_args->force_reset, "Reset Elasticsearch mappings and settings. "