Implemented batch ingest API and tests

include/amduat/tgk/store.h

@@ -89,6 +89,10 @@ typedef struct {
   bool (*ingest_artifact)(void *ctx,
                           amduat_reference_t ref,
                           amduat_artifact_t artifact);
+  bool (*ingest_batch)(void *ctx,
+                       const amduat_reference_t *refs,
+                       const amduat_artifact_t *artifacts,
+                       size_t len);
   bool (*remove_artifact)(void *ctx, amduat_reference_t ref);
   amduat_tgk_graph_error_t (*resolve_edge)(void *ctx,
                                            amduat_reference_t ref,
@@ -138,6 +142,11 @@ bool amduat_tgk_store_ingest_artifact(amduat_tgk_store_t *store,
                                       amduat_reference_t ref,
                                       amduat_artifact_t artifact);
 
+bool amduat_tgk_store_ingest_batch(amduat_tgk_store_t *store,
+                                   const amduat_reference_t *refs,
+                                   const amduat_artifact_t *artifacts,
+                                   size_t len);
+
 bool amduat_tgk_store_remove_artifact(amduat_tgk_store_t *store,
                                       amduat_reference_t ref);
 

src/adapters/tgk_store_fs/tgk_store_fs.c

@@ -145,6 +145,20 @@ static bool amduat_tgk_store_fs_ingest_artifact(void *ctx,
   return mem_ops.ingest_artifact(&fs->mem, ref, artifact);
 }
 
+static bool amduat_tgk_store_fs_ingest_batch(
+    void *ctx,
+    const amduat_reference_t *refs,
+    const amduat_artifact_t *artifacts,
+    size_t len) {
+  amduat_tgk_store_fs_t *fs = (amduat_tgk_store_fs_t *)ctx;
+  amduat_tgk_store_ops_t mem_ops = amduat_tgk_store_mem_ops();
+
+  if (fs == NULL || mem_ops.ingest_batch == NULL) {
+    return false;
+  }
+  return mem_ops.ingest_batch(&fs->mem, refs, artifacts, len);
+}
+
 static bool amduat_tgk_store_fs_remove_artifact(void *ctx,
                                                 amduat_reference_t ref) {
   amduat_tgk_store_fs_t *fs = (amduat_tgk_store_fs_t *)ctx;
@@ -385,6 +399,7 @@ amduat_tgk_store_ops_t amduat_tgk_store_fs_ops(void) {
   ops.get_config = amduat_tgk_store_fs_get_config;
   ops.snapshot_id = amduat_tgk_store_fs_snapshot_id;
   ops.ingest_artifact = amduat_tgk_store_fs_ingest_artifact;
+  ops.ingest_batch = amduat_tgk_store_fs_ingest_batch;
   ops.remove_artifact = amduat_tgk_store_fs_remove_artifact;
   ops.resolve_edge = amduat_tgk_store_fs_resolve_edge;
   ops.edges_from = amduat_tgk_store_fs_edges_from;

src/adapters/tgk_store_mem/tgk_store_mem.c

@@ -694,60 +694,138 @@ static bool amduat_tgk_store_mem_snapshot_id(
   return true;
 }
 
-static bool amduat_tgk_store_mem_ingest_artifact(
-    void *ctx,
-    amduat_reference_t ref,
-    amduat_artifact_t artifact) {
-  amduat_tgk_store_mem_t *mem = (amduat_tgk_store_mem_t *)ctx;
-  size_t i;
-  size_t insert_at;
-  amduat_tgk_store_mem_artifact_t entry;
+static bool amduat_tgk_store_mem_ingest_batch_impl(
+    amduat_tgk_store_mem_t *mem,
+    const amduat_reference_t *refs,
+    const amduat_artifact_t *artifacts,
+    size_t len) {
+  amduat_tgk_store_mem_artifact_t *pending = NULL;
+  size_t pending_len = 0;
+  size_t pending_cap = 0;
   amduat_tgk_graph_edge_view_t *edges = NULL;
   size_t edges_len = 0;
+  size_t i;
+  size_t j;
 
   if (mem == NULL) {
     return false;
   }
-  if (!amduat_tgk_store_mem_hash_id_supported(mem, ref.hash_id)) {
-    return false;
+  if (len == 0) {
+    return true;
   }
-  if (artifact.bytes.len != 0 && artifact.bytes.data == NULL) {
+  if (refs == NULL || artifacts == NULL) {
     return false;
   }
 
-  for (i = 0; i < mem->artifacts_len; ++i) {
-    if (!amduat_reference_eq(mem->artifacts[i].ref, ref)) {
+  for (i = 0; i < len; ++i) {
+    amduat_reference_t ref = refs[i];
+    amduat_artifact_t artifact = artifacts[i];
+    bool seen_duplicate = false;
+
+    if (!amduat_tgk_store_mem_hash_id_supported(mem, ref.hash_id)) {
+      goto cleanup;
+    }
+    if (artifact.bytes.len != 0 && artifact.bytes.data == NULL) {
+      goto cleanup;
+    }
+
+    for (j = 0; j < mem->artifacts_len; ++j) {
+      if (!amduat_reference_eq(mem->artifacts[j].ref, ref)) {
+        continue;
+      }
+      if (!amduat_tgk_store_mem_artifact_value_eq(&mem->artifacts[j],
+                                                  artifact)) {
+        goto cleanup;
+      }
+      seen_duplicate = true;
+      break;
+    }
+    if (seen_duplicate) {
       continue;
     }
-    if (amduat_tgk_store_mem_artifact_value_eq(&mem->artifacts[i], artifact)) {
-      return true;
+
+    for (j = 0; j < pending_len; ++j) {
+      if (!amduat_reference_eq(pending[j].ref, ref)) {
+        continue;
+      }
+      if (!amduat_tgk_store_mem_artifact_value_eq(&pending[j], artifact)) {
+        goto cleanup;
+      }
+      seen_duplicate = true;
+      break;
     }
-    return false;
+    if (seen_duplicate) {
+      continue;
+    }
+
+    if (pending_len == pending_cap) {
+      size_t new_cap = (pending_cap == 0) ? 4u : pending_cap * 2u;
+      amduat_tgk_store_mem_artifact_t *next =
+          (amduat_tgk_store_mem_artifact_t *)realloc(
+              pending, new_cap * sizeof(*pending));
+      if (next == NULL) {
+        goto cleanup;
+      }
+      pending = next;
+      pending_cap = new_cap;
+    }
+
+    if (!amduat_tgk_store_mem_artifact_clone(ref, artifact,
+                                             &pending[pending_len])) {
+      goto cleanup;
+    }
+    pending_len++;
   }
 
-  if (mem->artifacts_len == mem->artifacts_cap) {
-    size_t new_cap = (mem->artifacts_cap == 0) ? 8u : mem->artifacts_cap * 2u;
+  if (pending_len == 0) {
+    free(pending);
+    return true;
+  }
+
+  if (mem->artifacts_len > SIZE_MAX - pending_len) {
+    goto cleanup;
+  }
+  if (mem->artifacts_len + pending_len > mem->artifacts_cap) {
+    size_t new_cap = mem->artifacts_cap == 0 ? pending_len
+                                             : mem->artifacts_cap;
+    while (new_cap < mem->artifacts_len + pending_len) {
+      if (new_cap > SIZE_MAX / 2) {
+        goto cleanup;
+      }
+      new_cap *= 2u;
+    }
+    if (new_cap < mem->artifacts_len + pending_len) {
+      new_cap = mem->artifacts_len + pending_len;
+    }
     amduat_tgk_store_mem_artifact_t *next =
         (amduat_tgk_store_mem_artifact_t *)realloc(
             mem->artifacts, new_cap * sizeof(*mem->artifacts));
     if (next == NULL) {
-      return false;
+      goto cleanup;
     }
     mem->artifacts = next;
     mem->artifacts_cap = new_cap;
   }
 
-  insert_at = mem->artifacts_len;
-  if (!amduat_tgk_store_mem_artifact_clone(ref, artifact, &entry)) {
-    return false;
-  }
-  mem->artifacts[insert_at] = entry;
-  mem->artifacts_len++;
+  {
+    size_t appended_len = pending_len;
+    size_t appended_start = mem->artifacts_len;
 
-  if (!amduat_tgk_store_mem_edges_build(mem, &edges, &edges_len)) {
-    mem->artifacts_len--;
-    amduat_tgk_store_mem_artifact_free(&mem->artifacts[insert_at]);
-    return false;
+    memcpy(mem->artifacts + appended_start, pending,
+           appended_len * sizeof(*pending));
+    mem->artifacts_len += appended_len;
+    free(pending);
+    pending = NULL;
+    pending_len = 0;
+
+    if (!amduat_tgk_store_mem_edges_build(mem, &edges, &edges_len)) {
+      for (i = 0; i < appended_len; ++i) {
+        amduat_tgk_store_mem_artifact_free(
+            &mem->artifacts[appended_start + i]);
+      }
+      mem->artifacts_len = appended_start;
+      return false;
+    }
   }
 
   amduat_tgk_store_mem_edge_view_list_free(mem->edges, mem->edges_len);
@@ -755,6 +833,32 @@ static bool amduat_tgk_store_mem_ingest_artifact(
   mem->edges_len = edges_len;
   mem->snapshot_id++;
   return true;
+
+cleanup:
+  for (i = 0; i < pending_len; ++i) {
+    amduat_tgk_store_mem_artifact_free(&pending[i]);
+  }
+  free(pending);
+  return false;
+}
+
+static bool amduat_tgk_store_mem_ingest_artifact(
+    void *ctx,
+    amduat_reference_t ref,
+    amduat_artifact_t artifact) {
+  amduat_tgk_store_mem_t *mem = (amduat_tgk_store_mem_t *)ctx;
+
+  return amduat_tgk_store_mem_ingest_batch_impl(mem, &ref, &artifact, 1);
+}
+
+static bool amduat_tgk_store_mem_ingest_batch(
+    void *ctx,
+    const amduat_reference_t *refs,
+    const amduat_artifact_t *artifacts,
+    size_t len) {
+  amduat_tgk_store_mem_t *mem = (amduat_tgk_store_mem_t *)ctx;
+
+  return amduat_tgk_store_mem_ingest_batch_impl(mem, refs, artifacts, len);
 }
 
 static bool amduat_tgk_store_mem_remove_artifact(void *ctx,
@@ -1187,6 +1291,7 @@ amduat_tgk_store_ops_t amduat_tgk_store_mem_ops(void) {
   ops.get_config = amduat_tgk_store_mem_get_config;
   ops.snapshot_id = amduat_tgk_store_mem_snapshot_id;
   ops.ingest_artifact = amduat_tgk_store_mem_ingest_artifact;
+  ops.ingest_batch = amduat_tgk_store_mem_ingest_batch;
   ops.remove_artifact = amduat_tgk_store_mem_remove_artifact;
   ops.resolve_edge = amduat_tgk_store_mem_resolve_edge;
   ops.edges_from = amduat_tgk_store_mem_edges_from;

src/tgk_stack/store/store.c

@@ -48,6 +48,16 @@ bool amduat_tgk_store_ingest_artifact(amduat_tgk_store_t *store,
   return store->ops.ingest_artifact(store->ctx, ref, artifact);
 }
 
+bool amduat_tgk_store_ingest_batch(amduat_tgk_store_t *store,
+                                   const amduat_reference_t *refs,
+                                   const amduat_artifact_t *artifacts,
+                                   size_t len) {
+  if (store == NULL || store->ops.ingest_batch == NULL) {
+    return false;
+  }
+  return store->ops.ingest_batch(store->ctx, refs, artifacts, len);
+}
+
 bool amduat_tgk_store_remove_artifact(amduat_tgk_store_t *store,
                                       amduat_reference_t ref) {
   if (store == NULL || store->ops.remove_artifact == NULL) {

tests/tgk/test_tgk_store_mem.c

@@ -918,6 +918,135 @@ cleanup:
   return exit_code;
 }
 
+static int test_ingest_batch_epoch(void) {
+  amduat_tgk_store_mem_t mem;
+  amduat_tgk_store_t store;
+  amduat_tgk_store_config_t config;
+  amduat_tgk_identity_domain_t domains[1];
+  uint32_t edge_tags[1];
+  amduat_tgk_edge_type_id_t edge_types[1];
+  amduat_asl_encoding_profile_id_t encodings[1];
+  amduat_reference_t refs[2];
+  amduat_artifact_t artifacts[2];
+  amduat_octets_t edge_bytes;
+  amduat_reference_t node_a;
+  amduat_reference_t node_b;
+  amduat_reference_t payload;
+  uint8_t digest_ref1[32];
+  uint8_t digest_ref2[32];
+  uint8_t digest_a[32];
+  uint8_t digest_b[32];
+  uint8_t digest_payload[32];
+  amduat_tgk_edge_body_t edge;
+  amduat_reference_t edge_from[1];
+  amduat_reference_t edge_to[1];
+  amduat_tgk_snapshot_id_t epoch_before = 0;
+  amduat_tgk_snapshot_id_t epoch_after = 0;
+  amduat_tgk_graph_scan_result_t scan;
+  int exit_code = 1;
+
+  memset(&mem, 0, sizeof(mem));
+  memset(&config, 0, sizeof(config));
+  memset(&edge, 0, sizeof(edge));
+  edge_bytes = amduat_octets(NULL, 0);
+
+  domains[0].encoding_profile = AMDUAT_ENC_ASL1_CORE_V1;
+  domains[0].hash_id = AMDUAT_HASH_ASL1_ID_SHA256;
+  edge_tags[0] = TYPE_TAG_TGK1_EDGE_V1;
+  edge_types[0] = 0x10;
+  encodings[0] = TGK1_EDGE_ENC_V1;
+
+  config.id_space.domains = domains;
+  config.id_space.domains_len = 1;
+  config.tgk_profiles.edge_tags = edge_tags;
+  config.tgk_profiles.edge_tags_len = 1;
+  config.tgk_profiles.edge_types = edge_types;
+  config.tgk_profiles.edge_types_len = 1;
+  config.tgk_profiles.encodings = encodings;
+  config.tgk_profiles.encodings_len = 1;
+
+  if (!amduat_tgk_store_mem_init(&mem, config, NULL, 0)) {
+    fprintf(stderr, "batch epoch init failed\n");
+    return 1;
+  }
+  amduat_tgk_store_init(&store, config, amduat_tgk_store_mem_ops(), &mem);
+
+  if (!amduat_tgk_store_snapshot_id(&store, &epoch_before)) {
+    fprintf(stderr, "batch epoch get failed\n");
+    goto cleanup_mem;
+  }
+
+  refs[0] = make_ref(0x81, digest_ref1);
+  refs[1] = make_ref(0x82, digest_ref2);
+  node_a = make_ref(0xa1, digest_a);
+  node_b = make_ref(0xb1, digest_b);
+  payload = make_ref(0xe1, digest_payload);
+
+  edge.type = 0x10;
+  edge_from[0] = node_a;
+  edge.from = edge_from;
+  edge.from_len = 1;
+  edge_to[0] = node_b;
+  edge.to = edge_to;
+  edge.to_len = 1;
+  edge.payload = payload;
+
+  if (!amduat_enc_tgk1_edge_encode_v1(&edge, &edge_bytes)) {
+    fprintf(stderr, "batch epoch encode failed\n");
+    goto cleanup_mem;
+  }
+
+  artifacts[0] = amduat_artifact_with_type(
+      edge_bytes, amduat_type_tag(TYPE_TAG_TGK1_EDGE_V1));
+  artifacts[1] = artifacts[0];
+
+  if (!amduat_tgk_store_ingest_batch(&store, refs, artifacts, 2)) {
+    fprintf(stderr, "batch ingest failed\n");
+    goto cleanup_mem;
+  }
+  if (!amduat_tgk_store_snapshot_id(&store, &epoch_after)) {
+    fprintf(stderr, "batch epoch after ingest failed\n");
+    goto cleanup_mem;
+  }
+  if (epoch_after != epoch_before + 1u) {
+    fprintf(stderr, "batch epoch did not increment\n");
+    goto cleanup_mem;
+  }
+
+  if (!amduat_tgk_store_scan_edges(&store,
+                                   (amduat_tgk_edge_type_filter_t){0},
+                                   amduat_octets(NULL, 0), false, &scan)) {
+    fprintf(stderr, "batch scan failed\n");
+    goto cleanup_mem;
+  }
+  if (scan.edges.len != 2) {
+    fprintf(stderr, "batch scan count mismatch\n");
+    amduat_tgk_graph_scan_result_free(&scan);
+    goto cleanup_mem;
+  }
+  amduat_tgk_graph_scan_result_free(&scan);
+
+  if (!amduat_tgk_store_ingest_batch(&store, refs, artifacts, 2)) {
+    fprintf(stderr, "batch re-ingest failed\n");
+    goto cleanup_mem;
+  }
+  if (!amduat_tgk_store_snapshot_id(&store, &epoch_after)) {
+    fprintf(stderr, "batch epoch after re-ingest failed\n");
+    goto cleanup_mem;
+  }
+  if (epoch_after != epoch_before + 1u) {
+    fprintf(stderr, "batch epoch changed on no-op\n");
+    goto cleanup_mem;
+  }
+
+  exit_code = 0;
+
+cleanup_mem:
+  amduat_tgk_store_mem_free(&mem);
+  free((void *)edge_bytes.data);
+  return exit_code;
+}
+
 static int test_resolve_edge_unsupported(const test_env_t *env) {
   amduat_reference_t ref;
   amduat_tgk_edge_body_t body;
@@ -1134,6 +1263,7 @@ int main(void) {
       test_duplicate_edge_ref_conflict() != 0 ||
       test_scan_edges_pagination() != 0 ||
       test_ingest_remove_epoch() != 0 ||
+      test_ingest_batch_epoch() != 0 ||
       test_type_filter(&env) != 0 ||
       test_ordering(&env) != 0 ||
       test_adjacency(&env) != 0 ||