--- /dev/null
+/**
+ * GStreamer
+ * Copyright (C) 2005 Thomas Vander Stichele <thomas@apestaart.org>
+ * Copyright (C) 2005 Ronald S. Bultje <rbultje@ronald.bitfreak.net>
+ * Copyright (C) 2018 Jijoong Moon <jijoong.moon@samsung.com>
+ *
+ *
+ * This library is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Library General Public
+ * License as published by the Free Software Foundation; either
+ * version 2 of the License, or (at your option) any later version.
+ *
+ * This library is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Library General Public License for more details.
+ *
+ */
+/**
+ * @file gsttensormerge.c
+ * @date 03 July 2018
+ * @brief GStreamer plugin to merge tensors (as a filter for other general neural network filters)
+ * @see https://github.com/nnsuite/nnstreamer
+ * @author Jijoong Moon <jijoong.moon@samsung.com>
+ * @bug No known bugs except for NYI items
+ *
+ */
+
+/**
+ * SECTION:element-tensormerge
+ *
+ * A Merger that merge tensor stream to tensor stream for NN frameworks.
+ * The output is always in the format of other/tensor
+ *
+ * <refsect2>
+ * <title>Example launch line</title>
+ * |[
+ * gst-launch -v -m tensor_merge name=merge ! fakesink
+ * filesrc location=b.png ! pngdec ! videoscale ! imagefreeze ! videoconvert ! video/x-raw,format=RGB,width=100,height=100,framerate=0/1 ! tensor_converter ! merge.sink_0
+ * filesrc location=b.png ! pngdec ! videoscale ! imagefreeze ! videoconvert ! video/x-raw,format=RGB,width=100,height=100,framerate=0/1 ! tensor_converter ! merge.sink_1
+ * filesrc location=b.png ! pngdec ! videoscale ! imagefreeze ! videoconvert ! video/x-raw,format=RGB,width=100,height=100,framerate=0/1 ! tensor_converter ! merge.sink_2
+ * ]|
+ *
+ * |[
+ * gst-launch -v -m tensor_merge name=merge ! filesink location=merge.log
+ * multifilesrc location="testsequence_%1d.png" index=0 caps="image/png, framerate=(fraction)30/1" ! pngdec ! tensor_converter ! merge.sink_0
+ * multifilesrc location="testsequence_%1d.png" index=0 caps="image/png, framerate=(fraction)30/1" ! pngdec ! tensor_converter ! merge.sink_1
+ * multifilesrc location="testsequence_%1d.png" index=0 caps="image/png, framerate=(fraction)30/1" ! pngdec ! tensor_converter ! merge.sink_2
+ *
+ * </refsect2>
+ *
+ */
+
+
+#ifdef HAVE_CONFIG_H
+#include <config.h>
+#endif
+
+#include <string.h>
+#include <gst/gst.h>
+#include <glib.h>
+
+#include "gsttensormerge.h"
+
+GST_DEBUG_CATEGORY_STATIC (gst_tensor_merge_debug);
+#define GST_CAT_DEFAULT gst_tensor_merge_debug
+
+enum
+{
+ PROP_0,
+ PROP_MODE,
+ PROP_OPTION,
+ PROP_SILENT,
+};
+
+/**
+ * @brief the capabilities of the inputs and outputs.
+ * describe the real formats here.
+ */
+static GstStaticPadTemplate src_templ = GST_STATIC_PAD_TEMPLATE ("src",
+ GST_PAD_SRC,
+ GST_PAD_ALWAYS,
+ GST_STATIC_CAPS (GST_TENSOR_CAP_DEFAULT)
+ );
+
+static GstStaticPadTemplate sink_templ = GST_STATIC_PAD_TEMPLATE ("sink_%u",
+ GST_PAD_SINK,
+ GST_PAD_REQUEST,
+ GST_STATIC_CAPS (GST_TENSOR_CAP_DEFAULT)
+ );
+
+static void gst_tensor_merge_finalize (GObject * object);
+
+static gboolean gst_tensor_merge_handle_src_event (GstPad * pad,
+ GstObject * parent, GstEvent * event);
+static GstPad *gst_tensor_merge_request_new_pad (GstElement * element,
+ GstPadTemplate * templ, const gchar * name, const GstCaps * caps);
+static GstStateChangeReturn gst_tensor_merge_change_state (GstElement * element,
+ GstStateChange transition);
+static gboolean gst_tensor_merge_sink_event (GstCollectPads * pads,
+ GstCollectData * data, GstEvent * event, GstTensorMerge * tensor_merge);
+static GstFlowReturn gst_tensor_merge_collected (GstCollectPads * pads,
+ GstTensorMerge * tensor_merge);
+
+static void gst_tensor_merge_set_property (GObject * object, guint prop_id,
+ const GValue * value, GParamSpec * pspec);
+static void gst_tensor_merge_get_property (GObject * object, guint prop_id,
+ GValue * value, GParamSpec * pspec);
+
+#define gst_tensor_merge_parent_class parent_class
+G_DEFINE_TYPE (GstTensorMerge, gst_tensor_merge, GST_TYPE_ELEMENT);
+
+/**
+ * @brief initialize the tensor_merge's class
+ */
+static void
+gst_tensor_merge_class_init (GstTensorMergeClass * klass)
+{
+ GObjectClass *gobject_class;
+ GstElementClass *gstelement_class;
+
+ gobject_class = (GObjectClass *) klass;
+ gstelement_class = (GstElementClass *) klass;
+
+ parent_class = g_type_class_peek_parent (klass);
+
+ gobject_class->finalize = gst_tensor_merge_finalize;
+ gobject_class->get_property = gst_tensor_merge_get_property;
+ gobject_class->set_property = gst_tensor_merge_set_property;
+
+ g_object_class_install_property (gobject_class, PROP_SILENT,
+ g_param_spec_boolean ("silent", "Silent", "Produce verbose output ?",
+ TRUE, G_PARAM_READWRITE | G_PARAM_STATIC_STRINGS));
+
+ g_object_class_install_property (gobject_class, PROP_MODE,
+ g_param_spec_string ("mode", "Mode", "Tensor transform mode ?",
+ "", G_PARAM_READWRITE));
+ g_object_class_install_property (gobject_class, PROP_OPTION,
+ g_param_spec_string ("option", "Option",
+ "Option for the tensor transform mode ?", "", G_PARAM_READWRITE));
+
+ gstelement_class->request_new_pad =
+ GST_DEBUG_FUNCPTR (gst_tensor_merge_request_new_pad);
+ gstelement_class->change_state =
+ GST_DEBUG_FUNCPTR (gst_tensor_merge_change_state);
+
+ gst_element_class_add_pad_template (gstelement_class,
+ gst_static_pad_template_get (&sink_templ));
+ gst_element_class_add_pad_template (gstelement_class,
+ gst_static_pad_template_get (&src_templ));
+
+ gst_element_class_set_details_simple (gstelement_class,
+ "TensorMerge",
+ "Merger/Tensor",
+ "Merge multiple tensor stream to tensor stream",
+ "Jijoong Moon <jijoong.moon@samsung.com>");
+
+}
+
+/**
+ * @brief initialize the new element
+ * instantiate pads and add them to element
+ * set pad calback functions
+ * initialize instance structure
+ */
+static void
+gst_tensor_merge_init (GstTensorMerge * tensor_merge)
+{
+ GstElementClass *klass = GST_ELEMENT_GET_CLASS (tensor_merge);
+
+ tensor_merge->srcpad =
+ gst_pad_new_from_template (gst_element_class_get_pad_template (klass,
+ "src"), "src");
+ gst_pad_set_event_function (tensor_merge->srcpad,
+ gst_tensor_merge_handle_src_event);
+
+ gst_element_add_pad (GST_ELEMENT (tensor_merge), tensor_merge->srcpad);
+
+ tensor_merge->collect = gst_collect_pads_new ();
+ gst_collect_pads_set_event_function (tensor_merge->collect,
+ (GstCollectPadsEventFunction)
+ GST_DEBUG_FUNCPTR (gst_tensor_merge_sink_event), tensor_merge);
+ gst_collect_pads_set_function (tensor_merge->collect,
+ (GstCollectPadsFunction) GST_DEBUG_FUNCPTR (gst_tensor_merge_collected),
+ tensor_merge);
+
+ tensor_merge->silent = TRUE;
+ gst_tensors_config_init (&tensor_merge->tensors_config);
+ tensor_merge->mode = GTT_END;
+ tensor_merge->option = NULL;
+ tensor_merge->loaded = FALSE;
+}
+
+static const gchar *gst_tensor_merge_mode_string[] = {
+ [GTT_LINEAR] = "linear",
+ [GTT_END] = "error",
+};
+
+static const gchar *gst_tensor_merge_linear_string[] = {
+ [LINEAR_WIDTH] = "width",
+ [LINEAR_HEIGHT] = "height",
+ [LINEAR_CHANNEL] = "channel",
+ [LINEAR_END] = NULL,
+};
+
+/**
+ * @brief Get the corresponding mode from the string value
+ * @param[in] str The string value for the mode
+ * @return corresponding mode for the string. GTT_END for errors
+ */
+static tensor_merge_mode
+gst_tensor_merge_get_mode (const gchar * str)
+{
+ int i;
+ for (i = 0; i < GTT_END; i++) {
+ if (!g_ascii_strcasecmp (gst_tensor_merge_mode_string[i], str))
+ return i;
+ }
+ return GTT_END;
+}
+
+/**
+ * @brief finalize vmethod
+ */
+static void
+gst_tensor_merge_finalize (GObject * object)
+{
+ GstTensorMerge *tensor_merge;
+ tensor_merge = GST_TENSOR_MERGE (object);
+
+ if (tensor_merge->collect)
+ gst_object_unref (tensor_merge->collect);
+ G_OBJECT_CLASS (parent_class)->finalize (object);
+}
+
+/**
+ * @brief making new request pad (gst element vmethod)
+ */
+static GstPad *
+gst_tensor_merge_request_new_pad (GstElement * element, GstPadTemplate * templ,
+ const gchar * req_name, const GstCaps * caps)
+{
+ GstPad *newpad;
+ GstTensorMerge *tensor_merge;
+ gchar *name;
+
+ g_return_val_if_fail (templ != NULL, NULL);
+ g_return_val_if_fail (GST_IS_TENSOR_MERGE (element), NULL);
+
+ tensor_merge = GST_TENSOR_MERGE (element);
+
+ name =
+ g_strdup_printf ("sink_%u",
+ tensor_merge->tensors_config.info.num_tensors);
+ newpad = gst_pad_new_from_template (templ, name);
+ g_free (name);
+
+ if (newpad) {
+
+ GstTensorMergePadData *tensormergepad;
+
+ tensormergepad = (GstTensorMergePadData *)
+ gst_collect_pads_add_pad (tensor_merge->collect, newpad,
+ sizeof (GstTensorMergePadData), NULL, TRUE);
+
+ tensormergepad->pad = newpad;
+ gst_pad_set_element_private (newpad, tensormergepad);
+ tensor_merge->tensors_config.info.num_tensors++;
+ gst_element_add_pad (element, newpad);
+ } else {
+ GST_WARNING_OBJECT (tensor_merge, "failed to create request pad");
+ }
+ return newpad;
+}
+
+/**
+ * @brief src event vmethod
+ */
+static gboolean
+gst_tensor_merge_handle_src_event (GstPad * pad, GstObject * parent,
+ GstEvent * event)
+{
+ GstEventType type;
+ type = event ? GST_EVENT_TYPE (event) : GST_EVENT_UNKNOWN;
+ switch (type) {
+ case GST_EVENT_SEEK:
+ return FALSE;
+ default:
+ break;
+ }
+
+ return gst_pad_event_default (pad, parent, event);
+}
+
+/**
+ * @brief sink event vmethod
+ */
+static gboolean
+gst_tensor_merge_sink_event (GstCollectPads * pads, GstCollectData * data,
+ GstEvent * event, GstTensorMerge * tensor_merge)
+{
+ gboolean ret;
+ switch (GST_EVENT_TYPE (event)) {
+ case GST_EVENT_FLUSH_STOP:
+ {
+ tensor_merge->need_segment = TRUE;
+ break;
+ }
+ default:
+ break;
+ }
+
+ ret = gst_collect_pads_event_default (pads, data, event, FALSE);
+ return ret;
+}
+
+/**
+ * @brief Compare dts & pts time and find earliest
+ * @param tensor_merge tensor merger
+ * @param old previous merge pad data
+ * @param new current merge pad data
+ * @return if > 0, new is earlier than old
+ */
+static gint
+gst_tensor_merge_compare_pads (GstTensorMerge * tensor_merge,
+ GstTensorMergePadData * old, GstTensorMergePadData * new)
+{
+ guint64 oldtime, newtime;
+ if (old == NULL || old->buffer == NULL)
+ return 1;
+ if (new == NULL || new->buffer == NULL)
+ return -1;
+ if (GST_CLOCK_TIME_IS_VALID (old->dts_timestamp) &&
+ GST_CLOCK_TIME_IS_VALID (new->dts_timestamp)) {
+ oldtime = old->dts_timestamp;
+ newtime = new->dts_timestamp;
+ } else {
+ oldtime = old->pts_timestamp;
+ newtime = new->pts_timestamp;
+ }
+
+ if (!GST_CLOCK_TIME_IS_VALID (oldtime))
+ return -1;
+ if (!GST_CLOCK_TIME_IS_VALID (newtime))
+ return 1;
+
+ if (newtime < oldtime)
+ return 1;
+ else if (newtime > oldtime)
+ return -1;
+
+ return 0;
+}
+
+/**
+ * @brief Generate TensorConfig with TensorsConfig
+ * @param tensor_merge tensor merger
+ * @param configs Tensors Config Data
+ * @param config Tensor Config Data
+ * @return true / false
+ */
+gboolean
+gst_merge_tensors_config (GstTensorMerge * tensor_merge,
+ GstTensorsConfig * configs, GstTensorConfig * config)
+{
+ gboolean ret = FALSE;
+ switch (tensor_merge->mode) {
+ case GTT_LINEAR:
+ {
+ switch (tensor_merge->data_linear.direction) {
+ /* TODO : Currently we merge channel only. Have to extend W,H direction */
+ case LINEAR_CHANNEL:
+ {
+ int i, width, height, channel, batch;
+ tensor_type type;
+ type = configs->info.info[0].type;
+ batch = configs->info.info[0].dimension[3];
+ width = configs->info.info[0].dimension[1];
+ height = configs->info.info[0].dimension[2];
+ channel = configs->info.info[0].dimension[0];
+
+ for (i = 1; i < configs->info.num_tensors; i++) {
+ if (type != configs->info.info[i].type)
+ GST_ELEMENT_ERROR (tensor_merge, CORE, NEGOTIATION, (NULL),
+ (NULL));
+ if (batch != configs->info.info[i].dimension[3])
+ GST_ELEMENT_ERROR (tensor_merge, CORE, NEGOTIATION, (NULL),
+ (NULL));
+ if (width != configs->info.info[i].dimension[1])
+ GST_ELEMENT_ERROR (tensor_merge, CORE, NEGOTIATION, (NULL),
+ (NULL));
+ if (height != configs->info.info[i].dimension[2])
+ GST_ELEMENT_ERROR (tensor_merge, CORE, NEGOTIATION, (NULL),
+ (NULL));
+ channel += configs->info.info[i].dimension[0];
+ }
+
+ config->info.type = type;
+ config->info.dimension[0] = width;
+ config->info.dimension[1] = height;
+ config->info.dimension[2] = channel;
+ config->info.dimension[3] = batch;
+ config->rate_d = configs->rate_d;
+ config->rate_n = configs->rate_n;
+ }
+ ret = TRUE;
+ break;
+ case LINEAR_WIDTH:
+ case LINEAR_HEIGHT:
+ default:
+ ret = FALSE;
+ }
+ }
+ break;
+ default:
+ ret = FALSE;
+ }
+
+ return ret;
+}
+
+/**
+ * @brief Looping to generete outbut buffer for srcpad
+ * @param tensor_merge tensor merger
+ * @param tensor_buf output buffer for srcpad
+ * @param pts_time earliest pts time (present timestamp)
+ * @param dts_time earliest dts time (decoding timestamp)
+ * @return isEOS boolean EOS ( End of Stream )
+ */
+static gboolean
+gst_tensor_merge_collect_buffer (GstTensorMerge * tensor_merge,
+ GstBuffer * tensors_buf, GstClockTime * pts_time, GstClockTime * dts_time)
+{
+ GSList *walk = NULL;
+ GstTensorMergePadData *bestpad = NULL;
+ GstMemory *mem;
+ gboolean isEOS = FALSE;
+ gint old_numerator = G_MAXINT;
+ gint old_denominator = G_MAXINT;
+ gint counting = 0;
+ GstTensorConfig config;
+
+ walk = tensor_merge->collect->data;
+
+ while (walk) {
+ GstCollectData *data = (GstCollectData *) walk->data;
+ GstTensorMergePadData *pad = (GstTensorMergePadData *) data;
+ GstCaps *caps = gst_pad_get_current_caps (pad->pad);
+ GstStructure *s = gst_caps_get_structure (caps, 0);
+
+ gst_tensor_config_from_structure (&config, s);
+ g_assert (gst_tensor_config_validate (&config));
+
+ if (config.rate_d < old_denominator)
+ old_denominator = config.rate_d;
+ if (config.rate_n < old_numerator)
+ old_numerator = config.rate_n;
+
+ gst_caps_unref (caps);
+
+ walk = g_slist_next (walk);
+ GstBuffer *buf = NULL;
+ buf = gst_collect_pads_pop (tensor_merge->collect, data);
+ if (buf && GST_BUFFER_PTS_IS_VALID (buf)) {
+ pad->pts_timestamp =
+ gst_segment_to_running_time (&data->segment, GST_FORMAT_TIME,
+ GST_BUFFER_PTS (buf));
+ } else {
+ pad->pts_timestamp = GST_CLOCK_TIME_NONE;
+ }
+ if (buf && GST_BUFFER_DTS_IS_VALID (buf)) {
+ pad->dts_timestamp =
+ gst_segment_to_running_time (&data->segment, GST_FORMAT_TIME,
+ GST_BUFFER_DTS (buf));
+ } else {
+ pad->dts_timestamp = GST_CLOCK_TIME_NONE;
+ }
+
+ pad->buffer = buf;
+ if (GST_IS_BUFFER (buf)) {
+ mem = gst_buffer_get_memory (buf, 0);
+ gst_buffer_append_memory (tensors_buf, mem);
+
+ if (pad->buffer != NULL) {
+ if (gst_tensor_merge_compare_pads (tensor_merge, bestpad, pad) > 0) {
+ bestpad = pad;
+ *pts_time = bestpad->pts_timestamp;
+ *dts_time = bestpad->dts_timestamp;
+ }
+ }
+
+ gst_buffer_unref (buf);
+ } else {
+ isEOS = TRUE;
+ }
+
+ tensor_merge->tensors_config.info.info[counting] = config.info;
+ counting++;
+ }
+
+ tensor_merge->tensors_config.rate_d = old_denominator;
+ tensor_merge->tensors_config.rate_n = old_numerator;
+
+ debug_print (!tensor_merge->silent, "pts %" GST_TIME_FORMAT,
+ GST_TIME_ARGS (*pts_time));
+ debug_print (!tensor_merge->silent, "dts %" GST_TIME_FORMAT,
+ GST_TIME_ARGS (*dts_time));
+
+ /* set timestamp */
+ GST_BUFFER_PTS (tensors_buf) = *pts_time;
+ GST_BUFFER_DTS (tensors_buf) = *dts_time;
+ return isEOS;
+}
+
+/**
+ * @brief Gst Collect Pads Function which is called once collect pads done.
+ * @param pads GstCollectPads
+ * @param tensor_merge Merger
+ * @return GstFlowReturn
+ */
+static GstFlowReturn
+gst_tensor_merge_collected (GstCollectPads * pads,
+ GstTensorMerge * tensor_merge)
+{
+ GstFlowReturn ret = GST_FLOW_OK;
+ GstBuffer *tensors_buf, *tensor_buf;
+ GstMemory *mem;
+ GstClockTime pts_time = GST_CLOCK_TIME_NONE;
+ GstClockTime dts_time = GST_CLOCK_TIME_NONE;
+ GstClockTime time = 0;
+ gboolean isEOS = FALSE;
+ GST_DEBUG_OBJECT (tensor_merge, " all pads are collected ");
+ if (tensor_merge->need_stream_start) {
+ gchar s_id[32];
+ g_snprintf (s_id, sizeof (s_id), " tensormerge - %08x ", g_random_int ());
+ gst_pad_push_event (tensor_merge->srcpad,
+ gst_event_new_stream_start (s_id));
+ tensor_merge->need_stream_start = FALSE;
+ }
+
+ tensors_buf = gst_buffer_new ();
+ tensor_buf = gst_buffer_new ();
+
+ isEOS =
+ gst_tensor_merge_collect_buffer (tensor_merge, tensors_buf, &pts_time,
+ &dts_time);
+
+ if (isEOS) {
+ if (tensors_buf)
+ gst_buffer_unref (tensors_buf);
+
+ if (tensor_buf)
+ gst_buffer_unref (tensor_buf);
+
+ gst_pad_push_event (tensor_merge->srcpad, gst_event_new_eos ());
+ ret = GST_FLOW_EOS;
+ goto beach;
+ }
+
+ if (!tensor_merge->negotiated) {
+ GstCaps *newcaps;
+ GstTensorConfig config;
+
+ if (gst_merge_tensors_config (tensor_merge, &tensor_merge->tensors_config,
+ &config)) {
+
+ g_assert (gst_tensor_config_validate (&config));
+ newcaps = gst_tensor_caps_from_config (&config);
+ } else {
+ goto nego_error;
+ }
+
+ if (!gst_pad_set_caps (tensor_merge->srcpad, newcaps)) {
+ gst_caps_unref (newcaps);
+ goto nego_error;
+ }
+
+ gst_caps_unref (newcaps);
+ tensor_merge->negotiated = TRUE;
+ }
+
+ if (tensor_merge->need_segment) {
+ GstSegment segment;
+
+ if (GST_CLOCK_TIME_IS_VALID (dts_time)) {
+ time = dts_time;
+ } else if (GST_CLOCK_TIME_IS_VALID (pts_time)) {
+ time = pts_time;
+ } else {
+ time = 0;
+ }
+
+ gst_segment_init (&segment, GST_FORMAT_TIME);
+ segment.start = time;
+ gst_pad_push_event (tensor_merge->srcpad, gst_event_new_segment (&segment));
+ tensor_merge->need_segment = FALSE;
+ }
+
+
+ /* TODO : Should handle Properly. This is just for the temperaly usage */
+ mem = gst_buffer_get_memory_range (tensors_buf, 0, -1);
+ gst_buffer_append_memory (tensor_buf, mem);
+ gst_buffer_copy_into (tensor_buf, tensors_buf, GST_BUFFER_COPY_TIMESTAMPS, 0,
+ -1);
+ gst_buffer_unref (tensors_buf);
+
+ ret = gst_pad_push (tensor_merge->srcpad, tensor_buf);
+ if (ret != GST_FLOW_OK) {
+ GST_WARNING_OBJECT (tensor_merge, "pushed outbuf, result = %s",
+ gst_flow_get_name (ret));
+ /* fall-through, returns result */
+ }
+beach:
+ return ret;
+nego_error:
+ {
+ GST_WARNING_OBJECT (tensor_merge, "failed to set caps");
+ GST_ELEMENT_ERROR (tensor_merge, CORE, NEGOTIATION, (NULL), (NULL));
+ return GST_FLOW_NOT_NEGOTIATED;
+ }
+}
+
+/**
+ * @brief Ready --> Pasuse State Change
+ */
+static void
+gst_tensor_merge_ready_to_paused (GstTensorMerge * tensor_merge)
+{
+ tensor_merge->need_stream_start = TRUE;
+ tensor_merge->need_segment = TRUE;
+ tensor_merge->negotiated = FALSE;
+ gst_collect_pads_start (tensor_merge->collect);
+}
+
+/**
+ * @brief change state (gst element vmethod)
+ */
+static GstStateChangeReturn
+gst_tensor_merge_change_state (GstElement * element, GstStateChange transition)
+{
+ GstTensorMerge *tensor_merge;
+ GstStateChangeReturn ret;
+ tensor_merge = GST_TENSOR_MERGE (element);
+ switch (transition) {
+ case GST_STATE_CHANGE_READY_TO_PAUSED:
+ gst_tensor_merge_ready_to_paused (tensor_merge);
+ break;
+ case GST_STATE_CHANGE_PAUSED_TO_READY:
+ gst_collect_pads_stop (tensor_merge->collect);
+ break;
+ default:
+ break;
+ }
+
+ ret = GST_ELEMENT_CLASS (parent_class)->change_state (element, transition);
+ if (ret == GST_STATE_CHANGE_FAILURE)
+ return ret;
+ switch (transition) {
+ case GST_STATE_CHANGE_PAUSED_TO_READY:
+ break;
+ default:
+ break;
+ }
+
+ return ret;
+}
+
+/**
+ * @brief Setup internal data (data_* in GstTensor_Merge)
+ * @param[in/out] filter "this" pointer. mode & option MUST BE set already.
+ */
+static void
+gst_tensor_merge_set_option_data (GstTensorMerge * filter)
+{
+ if (filter->mode == GTT_END || filter->option == NULL)
+ return;
+ switch (filter->mode) {
+ case GTT_LINEAR:
+ {
+ filter->data_linear.direction =
+ find_key_strv (gst_tensor_merge_linear_string, filter->option);
+ g_assert (filter->data_linear.direction >= 0);
+ filter->loaded = TRUE;
+ }
+ break;
+ default:
+ g_printerr ("Cannot identify mode\n");
+ g_assert (0);
+ }
+}
+
+/**
+ * @brief Get property (gst element vmethod)
+ */
+static void
+gst_tensor_merge_set_property (GObject * object, guint prop_id,
+ const GValue * value, GParamSpec * pspec)
+{
+ GstTensorMerge *filter = GST_TENSOR_MERGE (object);
+ switch (prop_id) {
+ case PROP_SILENT:
+ filter->silent = g_value_get_boolean (value);
+ break;
+ case PROP_MODE:
+ filter->mode = gst_tensor_merge_get_mode (g_value_get_string (value));
+ g_assert (filter->mode != GTT_END);
+ gst_tensor_merge_set_option_data (filter);
+ break;
+ case PROP_OPTION:
+ filter->option = g_value_dup_string (value);
+ gst_tensor_merge_set_option_data (filter);
+ break;
+ default:
+ G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
+ break;
+ }
+}
+
+/**
+ * @brief Get property (gst element vmethod)
+ */
+static void
+gst_tensor_merge_get_property (GObject * object, guint prop_id,
+ GValue * value, GParamSpec * pspec)
+{
+ GstTensorMerge *filter = GST_TENSOR_MERGE (object);
+ switch (prop_id) {
+ case PROP_SILENT:
+ g_value_set_boolean (value, filter->silent);
+ break;
+ case PROP_MODE:
+ g_value_set_string (value, gst_tensor_merge_mode_string[filter->mode]);
+ break;
+ case PROP_OPTION:
+ g_value_set_string (value, filter->option);
+ break;
+ default:
+ G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
+ break;
+ }
+}
+
+
+/**
+ * PACKAGE: this is usually set by autotools depending on some _INIT macro
+ * in configure.ac and then written into and defined in config.h, but we can
+ * just set it ourselves here in case someone doesn't use autotools to
+ * compile this code. GST_PLUGIN_DEFINE needs PACKAGE to be defined.
+ */
+#ifndef PACKAGE
+#define PACKAGE "tensor_merge"
+#endif
+
+/**
+ * @brief entry point to initialize the plug-in
+ * initialize the plug-in itself
+ * register the element factories and other features
+ */
+gboolean
+gst_tensor_merge_plugin_init (GstPlugin * tensormerge)
+{
+ /** debug category for fltering log messages
+ * exchange the string 'Template tensor_merge' with your description
+ */
+ GST_DEBUG_CATEGORY_INIT (gst_tensor_merge_debug, "tensor_merge", 0,
+ "Tensor Merger");
+ return gst_element_register (tensormerge, "tensor_merge",
+ GST_RANK_NONE, GST_TYPE_TENSOR_MERGE);
+}
+
+/**
+ * @brief gstreamer looks for this structure to register tensormerge
+ */
+GST_PLUGIN_DEFINE (GST_VERSION_MAJOR,
+ GST_VERSION_MINOR,
+ tensor_merge,
+ "tensor merge plugin",
+ gst_tensor_merge_plugin_init, VERSION, "LGPL", "GStreamer",
+ "http://gstreamer.net/");
projects / platform / upstream / nnstreamer.git / commitdiff