{"id":2911,"date":"2025-10-21T01:21:46","date_gmt":"2025-10-21T01:21:46","guid":{"rendered":"https:\/\/blog.samarthya.me\/wps\/?p=2911"},"modified":"2025-10-21T01:21:48","modified_gmt":"2025-10-21T01:21:48","slug":"building-a-minimal-langchain-ollama-demo","status":"publish","type":"post","link":"https:\/\/blog.samarthya.me\/wps\/2025\/10\/21\/building-a-minimal-langchain-ollama-demo\/","title":{"rendered":"Building a Minimal LangChain + Ollama Demo"},"content":{"rendered":"\n
\"\"<\/figure>\n\n\n\n

This post walks through a tiny, practical example that uses LangChain<\/code> primitives and a local Ollama <\/code>server to run a prompt-based summarization. The project is intentionally small so you can learn the core ideas quickly: extracting prompts, initializing an LLM client, composing a chain, and running it locally.<\/p>\n\n\n\n

Repository used in this post:\u00a0llangchain-tutorial<\/a><\/code><\/p>\n\n\n\n

What you’ll learn<\/h2>\n\n\n\n
    \n
  • How to separate prompts from code for rapid iteration.<\/li>\n\n\n\n
  • How to initialize and call a local Ollama <\/code>server from Python using\u00a0ChatOllama<\/code>.<\/li>\n\n\n\n
  • The basics of composing a LangChain chain: PromptTemplate | LLM.<\/li>\n\n\n\n
  • How to structure code for clarity and reusability.<\/li>\n<\/ul>\n\n\n\n

    Why use a local runtime (Ollama)?<\/h2>\n\n\n\n

    Local runtimes like Ollama let you iterate quickly without network latency and without sending data to a cloud provider \u2014 useful for testing, privacy, and cost control. In this demo we use Ollama as the model host and ChatOllama<\/code> from the langchain_ollama<\/code> integration to talk to it.<\/p>\n\n\n\n

    Project walkthrough<\/h2>\n\n\n\n

    Files we care about:<\/p>\n\n\n\n

      \n
    • main.py<\/code>\u00a0\u2014 entrypoint and orchestrator<\/li>\n\n\n\n
    • prompts\/summary_template.txt<\/code>\u00a0\u2014 extracted prompt template<\/li>\n\n\n\n
    • README.md<\/code>\u00a0\u2014 run instructions and a Mermaid flowchart<\/li>\n<\/ul>\n\n\n\n

      Core idea<\/h3>\n\n\n\n

      main.py<\/code> follows a simple pattern:<\/p>\n\n\n\n

        \n
      1. Load environment variables (OLLAMA_HOST<\/code>,\u00a0OLLAMA_PORT<\/code>).<\/li>\n\n\n\n
      2. Load a prompt template from\u00a0prompts\/summary_template.txt<\/code>.<\/li>\n\n\n\n
      3. Initialize a\u00a0ChatOllama<\/code>\u00a0client using the host and port.<\/li>\n\n\n\n
      4. Build a chain by composing a PromptTemplate with the LLM (PromptTemplate(...) | llm<\/code>).<\/li>\n\n\n\n
      5. Invoke the chain with\u00a0chain.invoke(input={\"information\": information})<\/code>\u00a0and print the response.<\/li>\n<\/ol>\n\n\n\n

        Keeping the prompt in a file makes it easy for non-developers to tweak the prompt text without touching code. The helper functions in main.py<\/code> (load_prompt<\/code>, init_ollama_llm<\/code>, build_chain<\/code>) make each step testable and reusable.<\/p>\n\n\n\n

        Prompt template<\/h3>\n\n\n\n

        The prompt is intentionally simple and lives in prompts\/summary_template.txt<\/code>:<\/p>\n\n\n\n

        Given the information {information}, about the person I want you to create:\n    1. A short summary\n    2. Two interesting facts about them.\n<\/code><\/pre>\n\n\n\n
        Key snippet: building and invoking the chain<\/h3>\n\n\n\n
        # load prompt<\/em>\nprompt_text = load_prompt(\"prompts\/summary_template.txt\")\n\n# init llm<\/em>\nllm = init_ollama_llm()\n\n# compose chain<\/em>\nchain = build_chain(prompt_text, [\"information\"], llm)\n\n# invoke<\/em>\nresponse = chain.invoke(input={\"information\": information})\nprint(response.content)\n<\/code><\/pre>\n\n\n\n
        Note: The keys in the input<\/code> dict must match the input_variables<\/code> declared when creating the PromptTemplate. Passing a different key (e.g., \"info\"<\/code>) without updating the template will raise a missing-variable error.<\/p>\n\n\n\n
        How to run the demo<\/h2>\n\n\n\n
          \n
        1. Copy the env example and edit if necessary:<\/li>\n<\/ol>\n\n\n\n
          cp .env.example .env\n<\/code><\/pre>\n\n\n\n
            \n
          1. Create a virtual environment and install dependencies:<\/li>\n<\/ol>\n\n\n\n
            python -m venv .venv\nsource .venv\/bin\/activate\npip install -r requirements.txt\n<\/code><\/pre>\n\n\n\n
              \n
            1. Ensure Ollama is running and reachable (default:\u00a0localhost:11434<\/code>). You can check:<\/li>\n<\/ol>\n\n\n\n
              curl http:\/\/localhost:11434\/api\/models\ncurl http:\/\/localhost:11434\/api\/tags\n<\/code><\/pre>\n\n\n\n
                \n
              1. Run the example:<\/li>\n<\/ol>\n\n\n\n
                python main.py\n<\/code><\/pre>\n\n\n\n
                You should see a short printed summary and two facts derived from the information<\/code> sample in the repo.<\/p>\n\n\n\n
                Common pitfalls and debugging<\/h2>\n\n\n\n
                  \n
                • Prompt variable mismatch: ensure the dict keys you pass to\u00a0chain.invoke<\/code>\u00a0match the PromptTemplate’s\u00a0input_variables<\/code>.<\/li>\n\n\n\n
                • Ollama connectivity: if the server isn’t running,\u00a0curl<\/code>\u00a0the endpoints above and verify the host\/port.<\/li>\n<\/ul>\n\n\n\n
                  Next steps and experiments<\/h2>\n\n\n\n
                    \n
                  • Add memory to the chain to keep conversation context across invocations.<\/li>\n\n\n\n
                  • Swap\u00a0ChatOllama<\/code>\u00a0for another provider (OpenAI, Anthropic) to compare behavior.<\/li>\n\n\n\n
                  • Add unit tests: validate\u00a0load_prompt<\/code>\u00a0and use a mocked LLM to ensure the prompt is formatted correctly.<\/li>\n\n\n\n
                  • Create a small web UI to let users paste text and see summarization results.<\/li>\n<\/ul>\n\n\n\n
                    Conclusion<\/h2>\n\n\n\n
                    This project shows how to structure a small LangChain-based demo that talks to a local LLM runtime. The key takeaways are: keep prompts separate, build small reusable helpers, and use local runtimes for fast iteration.<\/p>\n","protected":false},"excerpt":{"rendered":"
                    This post walks through a tiny, practical example that uses LangChain primitives and a local Ollama server to run a prompt-based summarization. The project is intentionally small so you can learn the core ideas quickly: extracting prompts, initializing an LLM client, composing a chain, and running it locally. Repository used in this post:\u00a0llangchain-tutorial What you’ll […]<\/p>\n","protected":false},"author":2,"featured_media":2913,"comment_status":"open","ping_status":"open","sticky":false,"template":"elementor_header_footer","format":"standard","meta":{"_exactmetrics_skip_tracking":false,"_exactmetrics_sitenote_active":false,"_exactmetrics_sitenote_note":"","_exactmetrics_sitenote_category":0,"footnotes":""},"categories":[34],"tags":[350,351,327],"class_list":["post-2911","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-technical","tag-langchain","tag-llm","tag-ollama"],"_links":{"self":[{"href":"https:\/\/blog.samarthya.me\/wps\/wp-json\/wp\/v2\/posts\/2911","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/blog.samarthya.me\/wps\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/blog.samarthya.me\/wps\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/blog.samarthya.me\/wps\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/blog.samarthya.me\/wps\/wp-json\/wp\/v2\/comments?post=2911"}],"version-history":[{"count":1,"href":"https:\/\/blog.samarthya.me\/wps\/wp-json\/wp\/v2\/posts\/2911\/revisions"}],"predecessor-version":[{"id":2914,"href":"https:\/\/blog.samarthya.me\/wps\/wp-json\/wp\/v2\/posts\/2911\/revisions\/2914"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/blog.samarthya.me\/wps\/wp-json\/wp\/v2\/media\/2913"}],"wp:attachment":[{"href":"https:\/\/blog.samarthya.me\/wps\/wp-json\/wp\/v2\/media?parent=2911"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/blog.samarthya.me\/wps\/wp-json\/wp\/v2\/categories?post=2911"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/blog.samarthya.me\/wps\/wp-json\/wp\/v2\/tags?post=2911"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}