refactor: replace "compression" with "quantization" across codebase

The library quantizes values (reduces resolution via right-shift)
but does not compress data. The caller chooses the storage type.
Replaced all "compress/decompress" wording with "quantize/restore"
in NatSpec, README, AGENTS.md, copilot instructions, and tests.

Author: 0xferit

.github/copilot-instructions.md

@@ -2,7 +2,7 @@
 
 ## Project Overview
 
-`uint-quantization-lib` is a pure-function Solidity library for shift-based `uint256` compression.
+`uint-quantization-lib` is a pure-function Solidity library for shift-based `uint256` quantization.
 The core implementation is `UintQuantizationLib` in `src/UintQuantizationLib.sol`, built around
 `Quant` (a `uint16` value type that packs `discardedBitWidth` and `encodedBitWidth`).
 

AGENTS.md

@@ -4,7 +4,7 @@ This file provides guidance to Claude Code (claude.ai/code) when working with co
 
 ## Project Overview
 
-`uint-quantization-lib` is a pure-function Solidity library for shift-based `uint256` lossy compression. The core mechanism is floor quantization via right-shifting. A `Quant` value type packs `(discardedBitWidth, encodedBitWidth)` into a single `uint16`, making the compression scheme explicit and reusable. The recommended pattern is `immutable` + `create(discardedBitWidth, encodedBitWidth)`.
+`uint-quantization-lib` is a pure-function Solidity library for shift-based `uint256` quantization. The core mechanism is floor quantization via right-shifting. A `Quant` value type packs `(discardedBitWidth, encodedBitWidth)` into a single `uint16`, making the quantization scheme explicit and reusable. The recommended pattern is `immutable` + `create(discardedBitWidth, encodedBitWidth)`.
 
 ## Commands
 

README.md

@@ -5,7 +5,7 @@
 
 On-chain values routinely carry more resolution than the protocol needs, but storage charges for every bit you store, not every bit you use. Unnecessary resolution widens structs, fills extra slots, and costs 20,000 gas per cold write. You do not have to pay for resolution you do not use.
 
-This library quantizes `uint256` values via right-shift compression, packing more fields per storage slot and cutting gas on every write.
+This library quantizes `uint256` values via right-shift, packing more fields per storage slot and cutting gas on every write.
 
 **Quick start:**
 
@@ -14,8 +14,8 @@ import {Quant, UintQuantizationLib} from "uint-quantization-lib/src/UintQuantiza
 
 Quant private immutable SCHEME = UintQuantizationLib.create(32, 24);
 
-uint24 stored = uint24(SCHEME.encode(largeValue)); // compress
-uint256 restored = SCHEME.decode(stored); // decompress
+uint24 stored = uint24(SCHEME.encode(largeValue)); // quantize
+uint256 restored = SCHEME.decode(stored); // restore
 ```
 
 ## Installation
@@ -50,7 +50,7 @@ The `Quant` value type is a `uint16` with the following bit layout:
 | `q.encodedBitWidth()` | Bit-width of the encoded value (set at creation). |
 | `q.encode(value)` | Compresses `value` by discarding the low bits (floor). Reverts with `Overflow` if `value > max(q)`. |
 | `q.encode(value, true)` | Same as `encode(value)`, but also reverts with `NotAligned` if `value` is not step-aligned. |
-| `q.decode(encoded)` | Decompresses `encoded` back to the original scale. Discarded bits are restored as zeros (lower bound). |
+| `q.decode(encoded)` | Restores `encoded` back to the original scale. Discarded bits are restored as zeros (lower bound). |
 | `q.decodeMax(encoded)` | Like `decode`, but fills discarded bits with ones (upper bound within the step). |
 | `q.isValid()` | True if `q` satisfies the invariants enforced by `create`. Use to validate hand-wrapped `Quant` values. |
 | `q.fits(value)` | True if `value` fits within the scheme's representable range. |
@@ -81,7 +81,7 @@ contract StakingVault {
 
     mapping(address => uint96) internal stakes;
 
-    /// Floor-encodes msg.value and stores the compressed amount.
+    /// Floor-encodes msg.value and stores the quantized amount.
     function stake() external payable {
         require(SCHEME.fits(msg.value), "amount exceeds scheme max");
         stakes[msg.sender] = uint96(SCHEME.encode(msg.value));
@@ -154,7 +154,7 @@ Showcase contracts under `src/showcase/` use `UintQuantizationLib` and compare:
 
 - Real-life example (production-style ETH staking):
   raw path uses realistic packed fields by default (`uint128 amount`, `uint64` timestamps, `bool active`)
-  in `RawETHStakingShowcase`, while the quantized path further compresses stake amount into `uint96`
+  in `RawETHStakingShowcase`, while the quantized path further reduces stake amount into `uint96`
   in `QuantizedETHStakingShowcase`.
 - Extreme example (upper-bound packing showcase):
   raw path stores 12 full-width `uint256` values (`RawExtremePackingShowcase`),

src/UintQuantizationLib.sol

@@ -5,10 +5,10 @@ pragma solidity ^0.8.25;
  * @title UintQuantizationLib
  * @author [0xferit](https://github.com/0xferit)
  * @custom:security-contact ferit@cryptolab.net
- * @notice Pure-function library for shift-based uint256 compression using a bundled config type.
+ * @notice Pure-function library for shift-based uint256 quantization using a bundled config type.
  *
  *         The `Quant` value type packs a `(discardedBitWidth, encodedBitWidth)` scheme into a single `uint16`,
- *         allowing callers to define the compression config once and invoke methods on it.
+ *         allowing callers to define the quantization scheme once and invoke methods on it.
  *         Type layout (uint16):
  *           bits 0-7  → discardedBitWidth (LSBs discarded during encoding)
  *           bits 8-15 → encodedBitWidth   (bit-width of the encoded value)

test/UintQuantizationLib.t.sol

@@ -209,7 +209,7 @@ contract UintQuantizationLibSmokeTest is Test {
     }
 
     // -------------------------------------------------------------------------
-    // Boundary: discardedBitWidth == 0 (identity / no compression)
+    // Boundary: discardedBitWidth == 0 (identity / no quantization)
     // -------------------------------------------------------------------------
 
     function test_discardedBitWidth_zero_identity() public view {