Mastering Web Application Configurations A Journey Through Nginx, uWSGI, and Flask with Knowledge Graphs.

Join us for an engaging 25-minute session and explore the creation of knowledge graphs. My focus is on the art of gathering information and using it effectively.

This time, I'll specifically look at configuration files.

Do you face these challenges?

Unsure about how to write configuration files
Experimenting with sample configurations from websites results in errors
Encountering occasional errors in configurations
Facing issues after upgrading to a new version
Many of these problems often stem from inadequate configurations, leading to situations where things stop working due to a lack of updated knowledge.

How do you typically try to resolve them?

Seek advice from knowledgeable individuals
Conduct web searches
Ask ChatGPT
While these approaches might work, they often consume a significant amount of time.

Now, imagine using a knowledge graph in this scenario:

Your knowledge becomes visible
Faster issue resolution
Adaptability to version upgrades
Let's try creating a knowledge database using open-source tools. For example, envision setting up Python for web applications with Nginx + uWSGI + Flask. Each component has its unique setup, adding a layer of complexity. Let's tackle the challenge of finding the optimal configuration, especially in the occasionally perplexing world of uWSGI options.

This session is primarily designed for beginners. I'll provide straightforward explanations to make the content easily understandable.

Transcript

1. FOSSASIA 2024, Apr. 9 2024 14:20-14:45 Main Hall Mastering Web Application Configurations A Journey Through Nginx, uWSGI, and Flask with Knowledge Graphs. Koji Annoura ҊӜߒೋ koji@annoura.com 4PVSDF
   /FPK
   IUUQTOFPKDPNVTFDBTFTLOPXMFEHFHSBQI
2. Who am I ❖ Koji Annoura ❖ I live in Fukuoka Japan ❖ I love ❖ Apache Hop ❖ Data Orchestration App ❖ Apache User Group Japan ❖ I am a founder ❖ Neo4j Graph Database ❖ Neo4j Users Group Tokyo ❖ I am one of the founders. ❖ Graph Community MVP
3. I am studying English on Duolingo. ❖ Duolingo ❖ Over 3 years
4. Agenda ❖ 1. Web App ❖ Flask ❖ uWSGI ❖ 2. Let's write a conf fi le for uWSGI. ❖ Ask ChatGPT ❖ 3. One of the good solutions ❖ Knowledge Graph
5. u Web Server Gateway Interface (uWSGI) ❖ What is the correct pronunciation? ❖ 1. micro-whiskey ❖ 2. you-whizz-gee ❖ 3. mu-wiz-gee 4PVSDF
   /FPK
   IUUQTOFPKDPNVTFDBTFTLOPXMFEHFHSBQI
6. uWSGI ❖ micro-whiskey
7. 1. Web App
8. Introduction ❖ Who loves .conf, .ini fi les? ❖ Yes, Easy to setup ❖ No, Dif fi cult to setup ❖ No, what’s the best setting ❖ It’s like RPG ❖ You cannot achieve the goal in one attempt ❖ Seek hints from the master 4PVSDF
   8JLJQFEJB
   IUUQTFOXJLJQFEJBPSHXJLJ1PLÉNPO
9. Python Web App ❖ Which framework do you like? ❖ Web Frameworks ❖ Full-Stack ❖ Django ❖ Web2py ❖ Non Full-Stack ❖ Flask ❖ CherryPy ❖ FastAPI 4PVSDF
   1BMMFUT
   1SPKFDUT
   IUUQTQBMMFUTQSPKFDUTDPNQ fl BTL
10. Ruby Web App ❖ Please imagine that you are also working on a Ruby project. ❖ Web Frameworks ❖ Full-Stack ❖ Ruby on Rails 4PVSDF
    3VCZ
    PO
    3BJMT
    IUUQTSVCZPOSBJMTPSH
11. Java Web App ❖ Please imagine that you are also working on a Java project. ❖ Web Frameworks ❖ Spring Boot ❖ Play Framework 4PVSDF
    4QSJOH
    IUUQTTQSJOHJPQSPKFDUTTQSJOHCPPU
12. If you have 3 different lang Web App ❖ How do users access the web sites? ❖ Web Server? ❖ Python ❖ Flask run (port: 5000?) ❖ Ruby ❖ Webrick (port: 3000?) ❖ Java ❖ Tomcat (port: 8080?)
13. One Web Server ❖ This is a typical/traditional server con fi guration. ❖ Web Server ❖ Nginx ❖ Apache Web Server
14. uWSGI ❖ In the case of Flask, it's common to use uWSGI. ❖ uWSGI ❖ Connects Nginx and Flask
15. Traditional Problem-Solving Methods ❖ Traditional methods of resolving con fi guration issues: ❖ 1. Seeking advice from experts ❖ 2. Conducting web searches ❖ 3. Using AI assistants like ChatGPT 4PVSDF
    8JLJQFEJB
    IUUQTFOXJLJQFEJBPSHXJLJ1SPCMFN@TPMWJOH
16. 2. Let's write a conf file for uWSGI with ChatGPT
17. Teach me how to connect Nginx and Flask? ❖ The simplest connection ❖ ChatGPT ❖ To connect Nginx and a Flask app, you typically use Nginx as a reverse proxy server. 4PVSDF
    0QFO"*
    IUUQTDIBUPQFOBJDPN
18. How about using uWSGI? ❖ Common connections ❖ ChatGPT ❖ Using uWSGI with Nginx is a common setup for deploying Flask applications in production environments. 4PVSDF
    0QFO"*
    IUUQTDIBUPQFOBJDPN
19. .ini File for uWSGI ❖ Simple con fi guration fi le ❖ ChatGPT ❖ You can do this by creating a con fi guration fi le for uWSGI (e.g., app.ini) ɹ with theɹ following content. 4PVSDF
    0QFO"*
    IUUQTDIBUPQFOBJDPN
20. .ini File for uWSGI ❖ Simple con fi guration fi le ❖ ChatGPT ❖ You can do this by creating a con fi guration fi le for uWSGI (e.g., app.ini) ɹ with theɹ following content. 4PVSDF
    0QFO"*
    IUUQTDIBUPQFOBJDPN
21. How do I automatically reload when the app changes? ❖ Add option ❖ ChatGPT ❖ you can use uWSGI's --py-autoreload option. 4PVSDF
    0QFO"*
    IUUQTDIBUPQFOBJDPN
22. virtual environment (venv). ❖ ChatGPT ❖ If you're encountering issues with uWSGI's auto- reload feature while using a virtual environment (venv). 4PVSDF
    0QFO"*
    IUUQTDIBUPQFOBJDPN
23. Correct settings for virtual environment (venv)? ❖ ChatGPT ❖ If you're encountering issues with uWSGI's auto- reload feature while using a virtual environment (venv), it's likely due to the way uWSGI is detecting fi le changes within the virtual environment. 4PVSDF
    0QFO"*
    IUUQTDIBUPQFOBJDPN
24. Is virtualenv or venv correct than pythonpath. ❖ ChatGPT ❖ You're correct. Using virtualenv or venv to isolate the Python environment is indeed a more common and preferred approach compared to specifying the Python executable directly with pythonpath. 4PVSDF
    0QFO"*
    IUUQTDIBUPQFOBJDPN
25. What options should I add to my ini setup? ❖ ChatGPT ❖ Here's a list of commonly used options along with their descriptions: 4PVSDF
    0QFO"*
    IUUQTDIBUPQFOBJDPN
26. The “socket-backlog” uWSGI option is correct? ❖ ChatGPT ❖ The socket-backlog option in uWSGI determines the maximum number of connections that the operating system will queue for your application when the application is handling too many connections to keep up with. 4PVSDF
    0QFO"*
    IUUQTDIBUPQFOBJDPN
27. No “socket-backlog” uWSGI option ❖ Options ❖ uWSGI core
28. I can't find the “socket-backlog” option. ❖ ChatGPT ❖ I apologize for the confusion. You're correct; there isn't a socket-backlog option in uWSGI. 4PVSDF
    0QFO"*
    IUUQTDIBUPQFOBJDPN
29. Hallucination (artificial intelligence) ❖ Have you ever been troubled by incorrect answers from AI? ❖ Hallucination or arti fi cial hallucination is a response generated by AI which contains false or misleading information presented as fact. 4PVSDF
    8JLJQFEJB
    IUJUQTFOXJLJQFEJBPSHXJLJ)BMMVDJOBUJPO@ BSUJ fi DJBM@JOUFMMJHFODF
30. 3. One of the good solutions
31. How can one resolve hallucinations? ❖ Waiting for the next version. ❖ Look for another better AI. ❖ Stop using AI.
32. One of Solutions (LLM + Knowledge Graph) ❖ Return True Facts ❖ Neo4j’s knowledge graph grounds LLM responses in validated facts using retrieval augmented generation (RAG). Easily add and update data, indexes, and RAG sources with our developer friendly schema that never requires redesign. Graph structure makes fi nding factual answers to multi-hop questions possible through traversals that follow a breadcrumb trail of connections. 4PVSDF
    /FPK
    IUUQTOFPKDPNHFOFSBUJWFBJ
33. Knowledge Graphs? ❖ knowledge graph ❖ is a knowledge base that uses a graph- structured data model or topology to represent and operate on data. 4PVSDF
    8JLJQFEJB
    IUUQTFOXJLJQFEJBPSHXJLJ,OPXMFEHF@HSBQI
34. Graph Database ❖ 3 Components ❖ Nodes (Records) ❖ Relationships (Related tables) ❖ Properties (item) 4PVSDF
    %#&OHJOFT
    IUUQTECFOHJOFTDPNFOSBOLJOHHSBQI ECNT
35. (:Movie)—(:Person) ❖ Nodes ❖ Movie (38) ❖ Person (133) ❖ Relationship ❖ ACTED_IN ❖ DIRECTED ❖ PRODUCED ❖ WROTE ❖ FOLLOWD ❖ REVIEWED
36. Matrix, (:Movie)—(:Person)—(:Movie)—(:Person) ❖ Nodes ❖ Movie (38) ❖ Person (133) ❖ Relationship ❖ ACTED_IN ❖ DIRECTED ❖ PRODUCED ❖ WROTE ❖ FOLLOWD ❖ REVIEWED
37. Matrix, (:Movie)—(:Person) ❖ Nodes ❖ Movie (3) ❖ Person (8) ❖ Relationship ❖ ACTED_IN ❖ DIRECTED ❖ PRODUCED ❖ WROTE ❖ FOLLOWD ❖ REVIEWED
38. Matrix, (:Movie)—(:Person)—(:Movie)—(:Person) ❖ Nodes ❖ Movie (11) ❖ Person (44) ❖ Relationship ❖ ACTED_IN ❖ DIRECTED ❖ PRODUCED ❖ WROTE ❖ FOLLOWD ❖ REVIEWED
39. uWSGI Options ❖ Options ❖ 1, uWSGI core ❖ 2, plugin: airbrake ❖ 3, plugin: alarm_curl ❖ 4, plugin: alarm_speech ❖ 5, plugin: alarm_xmpp ❖ ... ❖ 123, plugin: xslt ❖ 124, plugin: zabbix ❖ 125, plugin: zergpool
40. uWSGI Options ❖ Options ❖ 1, uWSGI core ❖ 2, plugin: airbrake ❖ 3, plugin: alarm_curl ❖ 4, plugin: alarm_speech ❖ 5, plugin: alarm_xmpp ❖ ... ❖ 123, plugin: xslt ❖ 124, plugin: zabbix ❖ 125, plugin: zergpool
41. uWSGI Options to Neo4j Graph Database ❖ Nodes ❖ Category (1) ❖ Type (125) ❖ Option (1254) ❖ Relationship ❖ TYPE ❖ OPTION
42. uWSGI Options (1254) ❖ Nodes ❖ Category (1) ❖ Type (125) ❖ Option (1254) ❖ Relationship ❖ TYPE ❖ OPTION
43. uWSGI Options ❖ Nodes ❖ Category (1) ❖ Type (125) ❖ Option (1254) ❖ Relationship ❖ TYPE ❖ OPTION
44. Only the correct options are set w/Knoledge Graph ❖ LLM + Knowledge Graph 4PVSDF
    0QFO"*
    IUUQTDIBUPQFOBJDPN
45. You can add “Same as” relationships ❖ 4 diff option names ❖ plugin:python ❖ pyhome (-H) ❖ virtualenv (-H) ❖ home (-H) ❖ venv (-H) ❖ You can merge them into one option name.
46. You can add “Check it” relationships ❖ No logs ❖ disable-logging = true ❖ But you need some log ❖ log-4xx = true ❖ log-5xx = true ❖ log-ioerror = true
47. Ground LLMs with Knowledge Graph ❖ GenAI Apps ❖ https://neo4j.com/generativeai/ 4PVSDF
    /FPK
    IUUQTOFPKDPNHFOFSBUJWFBJ
48. Ground LLMs with Knowledge Graph ❖ Deep Partnerships with Generative AI Providers ❖ https://neo4j.com/generativeai/ 4PVSDF
    /FPK
    IUUQTOFPKDPNHFOFSBUJWFBJ
49. Conclusion ❖ The advancement of AI is remarkable. ❖ But ❖ Knowledge Graphs are particularly effective for representing facts. 4PVSDF
    /FPK
    IUUQTOFPKDPNHFOFSBUJWFBJ
50. Thanks ❖ Name ❖ Koji Annoura ❖ Linkedin ❖ kojiannoura ❖ X (Twitter) ❖ @kojiannoura