From 966f47b46f150d62e2f0683d6dfd2ac0b4f90c43 Mon Sep 17 00:00:00 2001
From: Cyber MacGeddon <cybermaggedon@gmail.com>
Date: Sat, 22 Mar 2025 00:26:00 +0000
Subject: [PATCH] - Template updates for pdf-ocr - Template updates for
 pdf-ocr-mistral

---
 .gitignore                               |  1 +
 templates/components.jsonnet             |  4 ++
 templates/components/mistral-ocr.jsonnet | 47 ++++++++++++++++++++++++
 templates/components/ocr.jsonnet         | 38 +++++++++++++++++++
 templates/values/images.jsonnet          |  1 +
 5 files changed, 91 insertions(+)
 create mode 100644 templates/components/mistral-ocr.jsonnet
 create mode 100644 templates/components/ocr.jsonnet

diff --git a/.gitignore b/.gitignore
index 357ecf1e..4d089211 100644
--- a/.gitignore
+++ b/.gitignore
@@ -9,6 +9,7 @@ trustgraph-base/trustgraph/base_version.py
 trustgraph-bedrock/trustgraph/bedrock_version.py
 trustgraph-embeddings-hf/trustgraph/embeddings_hf_version.py
 trustgraph-flow/trustgraph/flow_version.py
+trustgraph-ocr/trustgraph/ocr_version.py
 trustgraph-parquet/trustgraph/parquet_version.py
 trustgraph-vertexai/trustgraph/vertexai_version.py
 trustgraph-cli/trustgraph/
diff --git a/templates/components.jsonnet b/templates/components.jsonnet
index 121bd6a5..d0df569f 100644
--- a/templates/components.jsonnet
+++ b/templates/components.jsonnet
@@ -37,6 +37,10 @@
    "graph-rag": import "components/graph-rag.jsonnet",
    "document-rag": import "components/document-rag.jsonnet",
 
+   // OCR options
+   "ocr": import "components/ocr.jsonnet",
+   "mistral-ocr": import "components/mistral-ocr.jsonnet",
+
    // Librarian - document management
    "librarian": import "components/librarian.jsonnet",
 
diff --git a/templates/components/mistral-ocr.jsonnet b/templates/components/mistral-ocr.jsonnet
new file mode 100644
index 00000000..8049c514
--- /dev/null
+++ b/templates/components/mistral-ocr.jsonnet
@@ -0,0 +1,47 @@
+local images = import "values/images.jsonnet";
+local url = import "values/url.jsonnet";
+
+{
+
+    with:: function(key, value)
+        self + {
+            ["mistral-" + key]:: value,
+        },
+
+    "pdf-decoder" +: {
+    
+        create:: function(engine)
+
+            local envSecrets = engine.envSecrets("mistral-credentials")
+                .with_env_var("MISTRAL_TOKEN", "mistral-token");
+
+            local container =
+                engine.container("mistral-ocr")
+                    .with_image(images.trustgraph_flow)
+                    .with_command([
+                        "pdf-ocr-mistral",
+                        "-p",
+                        url.pulsar,
+                    ])
+                    .with_env_var_secrets(envSecrets)
+                    .with_limits("0.5", "128M")
+                    .with_reservations("0.1", "128M");
+
+            local containerSet = engine.containers(
+                "mistral-ocr", [ container ]
+            );
+
+            local service =
+                engine.internalService(containerSet)
+                .with_port(8080, 8080, "metrics");
+
+            engine.resources([
+                envSecrets,
+                containerSet,
+                service,
+            ])
+
+    },
+
+} + prompts
+
diff --git a/templates/components/ocr.jsonnet b/templates/components/ocr.jsonnet
new file mode 100644
index 00000000..4353b7f9
--- /dev/null
+++ b/templates/components/ocr.jsonnet
@@ -0,0 +1,38 @@
+local images = import "values/images.jsonnet";
+local url = import "values/url.jsonnet";
+
+{
+
+    "pdf-decoder" +: {
+    
+        create:: function(engine)
+
+            local container =
+                engine.container("pdf-ocr")
+                    .with_image(images.trustgraph_ocr)
+                    .with_command([
+                        "pdf-ocr",
+                        "-p",
+                        url.pulsar,
+                    ])
+                    .with_limits("1.0", "512M")
+                    .with_reservations("0.1", "512M");
+
+            local containerSet = engine.containers(
+                "pdf-ocr", [ container ]
+            );
+
+            local service =
+                engine.internalService(containerSet)
+                .with_port(8080, 8080, "metrics");
+
+            engine.resources([
+                envSecrets,
+                containerSet,
+                service,
+            ])
+
+    },
+
+} + prompts
+
diff --git a/templates/values/images.jsonnet b/templates/values/images.jsonnet
index dde235ce..54dbd016 100644
--- a/templates/values/images.jsonnet
+++ b/templates/values/images.jsonnet
@@ -11,6 +11,7 @@ local version = import "version.jsonnet";
     grafana: "docker.io/grafana/grafana:11.1.4",
     trustgraph_base: "docker.io/trustgraph/trustgraph-base:" + version,
     trustgraph_flow: "docker.io/trustgraph/trustgraph-flow:" + version,
+    trustgraph_ocr: "docker.io/trustgraph/trustgraph-ocr:" + version,
     trustgraph_bedrock: "docker.io/trustgraph/trustgraph-bedrock:" + version,
     trustgraph_vertexai: "docker.io/trustgraph/trustgraph-vertexai:" + version,
     trustgraph_hf: "docker.io/trustgraph/trustgraph-hf:" + version,