Add new model files and update Ansible playbooks for Qwen3.5 and Qwen3.6

- Created model files for Qwen3.5 with different context lengths.
- Added a new Ansible host variable file for CI server configuration.
- Updated GPU server host variables to include Ollama models and model files.
- Enhanced playbooks for Podman to manage Ollama models, including copying model files and checking existing models.
- Introduced a new playbook for setting up Qwen3.6 with specific configurations.
- Updated Windows SSH tasks to improve security and configuration management.
- Added a Python utility to parse Ollama model names from command output.
- Modified Terraform configurations to include new DNS entries and proxy settings.
- Improved Nginx proxy configurations with timeout settings.

bin/parse_ollama_models.py

@@ -0,0 +1,57 @@
+#!/usr/bin/env python3
+"""
+Utility to parse ollama ls output and extract model names.
+
+Replicates the Jinja2 filter logic from:
+  ollama_existing_models.stdout_lines
+  | reject('match', '^\\s*$|^(NAME|\\-+)')
+  | map('regex_replace', '^\\s*(\\S+)\\s.*', '\\1')
+  | list
+
+Usage:
+    python parse_ollama_models.py < input.txt
+    python parse_ollama_models.py models_output.txt
+    echo "model_name ..." | python parse_ollama_models.py
+"""
+
+import re
+import sys
+
+
+def parse_ollama_models(lines):
+    """Parse ollama ls output lines and return list of model names."""
+    results = []
+    for line in lines:
+        # Skip empty lines and header rows (NAME, ---)
+        if not line.strip() or re.match(r'^(NAME|\-+)$', line.strip()):
+            continue
+        # Extract the model name (first non-whitespace token)
+        match = re.match(r'^\s*(\S+)\s.*', line)
+        if match:
+            results.append(match.group(1))
+    return results
+
+
+def main():
+    # Read input from file argument, stdin, or skip
+    if len(sys.argv) > 1:
+        with open(sys.argv[1], 'r') as f:
+            lines = f.read().splitlines()
+    else:
+        lines = sys.stdin.read().splitlines()
+
+    models = parse_ollama_models(lines)
+
+    # Output: one model per line
+    for model in models:
+        print(model)
+
+    # Also write to a JSON file if a second arg is given
+    if len(sys.argv) > 2:
+        import json
+        with open(sys.argv[2], 'w') as f:
+            json.dump(models, f, indent=2)
+
+
+if __name__ == '__main__':
+    main()