docs: SEO overhaul — rank for LLM memory decay searches

[Neal006]

What This PR Does

Full discoverability push across Google and GitHub for the keywords: LLM memory decay, memory decay benchmark, LLM memory evaluation, RAG memory benchmark, temporal drift LLM, Ebbinghaus LLM memory.

---

GitHub repo metadata (already live)

• 20 topics set via gh repo edit: llm-memory, memory-decay, ebbinghaus, rag-evaluation, temporal-decay, conversation-memory, llm-benchmark, memory-evaluation, ai-evaluation, chatbot + 10 existing
• Description updated: all 5 backends + key differentiators in 160 chars

README.md — the main SEO asset

• ## How MemoryLens Compares table: head-to-head vs RAGAS, TruLens, DeepEval, MemGPT, LangChain ConversationBuffer — captures search traffic for competing tools
• Multi-seed results table (mean ± std, n=5 personas)
• Decay ablation table with academic references
• All 5 backends and 5 providers with free-tier flags
• Updated project structure tree with all new files
• Proper BibTeX citation block

CITATION.cff (new)

Standard software citation file — GitHub displays a "Cite this repository" panel in the sidebar when this file exists. Also indexed by Zenodo, OpenAlex, and Google Scholar.

pyproject.toml (new)

• 15 PyPI keywords for Python package search
• Optional provider extras: pip install memorylens[groq]
• Entry point: memorylens CLI command

docs/ (3 new files)

File	Target search query
why-memory-evaluation-matters.md	"why does LLM forget", "LLM memory decay explained"
comparison-with-existing-tools.md	"RAGAS vs MemoryLens", "LLM memory benchmark comparison"
adding-a-new-backend.md	"how to add LLM memory backend", "MemoryLens contribute"

Open issues (3 new searchable issue titles)

• #12 — EntityMemory backend
• #13 — EdTech domain scenario
• #14 — Qdrant/FAISS backend

🤖 Generated with Claude Code

[Copilot]

Pull request overview

This PR is a documentation/metadata overhaul aimed at improving discoverability for “LLM memory decay”–related searches, while also introducing initial Python packaging metadata (CITATION + pyproject.toml) to make the project more citable and installable.

Changes:

• Expands README.md and ROADMAP.md with SEO-targeted positioning, comparison tables, and updated benchmark/result narratives.
• Adds new long-form docs pages under docs/ to target specific search queries and guide contributions.
• Introduces CITATION.cff and pyproject.toml to support GitHub citation UX and Python packaging/entry points.

Reviewed changes

Copilot reviewed 10 out of 10 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
ROADMAP.md	Reframes roadmap around v0.3 shipped items and v0.4/v0.5 plans with SEO-forward wording
README.md	Major SEO/content expansion: results tables, comparisons, provider matrix, updated structure tree, citation block
pyproject.toml	Adds packaging metadata, dependencies, extras, and a CLI entry point
docs/why-memory-evaluation-matters.md	New explainer article targeting “why LLMs forget / memory decay” queries
docs/comparison-with-existing-tools.md	New comparison page vs RAGAS/TruLens/DeepEval/MemGPT/LangChain
docs/adding-a-new-backend.md	New contributor guide for adding memory backends (worked EntityMemory example)
CONTRIBUTING.md	Updates contributor onboarding and adds persona/scenario contribution guidance
CITATION.cff	Adds standard citation metadata for GitHub “Cite this repository” support
.github/ISSUE_TEMPLATE/new_backend.md	Expands backend issue template to capture research hypothesis + expected decay behavior
.github/ISSUE_TEMPLATE/feature_request.md	Expands feature request template to better fit benchmark/backend/scenario requests

Comments suppressed due to low confidence (1)

pyproject.toml:72

• The console script points at main:main, but main.py is a top-level module and isn’t included by the current setuptools config (only packages matching memory*, evaluation*, simulator*, utils* are discovered). This will make the installed memorylens command fail with ModuleNotFoundError: No module named 'main'. Either move the CLI into an importable package module (e.g. memorylens/cli.py) or explicitly include main as a py-modules entry / package it so the entry point resolves after installation.

[project.scripts]
memorylens = "main:main"

[project.urls]
Homepage = "https://github.com/Neal006/memorylens"
Repository = "https://github.com/Neal006/memorylens"
"Bug Tracker" = "https://github.com/Neal006/memorylens/issues"
Documentation = "https://github.com/Neal006/memorylens#readme"
Paper = "https://github.com/Neal006/memorylens/blob/main/paper/memorylens_paper.md"

[tool.setuptools.packages.find]
include = ["memory*", "evaluation*", "simulator*", "utils*"]

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.