x-iztro
中文文档:README.zh-CN.md
Give it a birth date and hour; get back a complete Zi Wei Dou Shu (紫微斗数, Purple Star Astrology) chart — twelve palaces, every star with its brightness and transformation, decadal and annual horoscopes — as typed objects in Rust, Python, or Go, plus a single call that renders the whole chart as text you can hand straight to an LLM.
What the LLM-ready output looks like
from x_iztro import Astro
astro = Astro()
chart = astro.by_solar("2000-8-16", 2, "female")
print(astro.astrolabe_to_prompt(chart))
=== 基本信息 ===
性别: 女
阳历: 2000-8-16
农历: 二〇〇〇年七月十七
干支: 庚辰 甲申 丙午 庚寅
时辰: 寅时 (03:00~05:00)
星座: 狮子座
生肖: 龙
命宫地支: 午
身宫地支: 戌
命主: 破军
身主: 文昌
五行局: 木三局
生年四化: 太阳禄, 武曲权, 太阴科, 天同忌
=== 十二宫 ===
--- 财帛 ---
天干地支: 戊寅
大限: 43-52
小限虚岁: 9, 21, 33, 45, 57, 69, 81, 93, 105, 117
十二神: 绝, 飞廉, 吊客, 岁驿
主星: 武曲(得)[权], 天相(庙)
辅星: 天马
杂耀: 解神, 三台, 天寿, 天巫, 天厨, 阴煞, 天哭
--- 夫妻 [来因] ---
天干地支: 庚辰
大限: 23-32
小限虚岁: 7, 19, 31, 43, 55, 67, 79, 91, 103, 115
十二神: 死, 将军, 岁建, 华盖
主星: 七杀(庙)
辅星: 右弼, 火星(陷)
杂耀: 封诰, 华盖
... (all twelve palaces)
The same text is available in six languages. Pass language="en-US" and the
same chart renders as:
=== Basic Info ===
Gender: female
Solar Date: 2000-8-16
Time: Tiger hour (03:00~05:00)
Soul Star: rebel
Five Elements Class: wood 3rd
Birth-Year Mutagen: sunA, generalB, moonC, fortunateD
--- wealth ---
Decadal: 43-52
Twelve Gods: dissipated, gossip, sorrowing, varied
Major Stars: general([+1])[B], minister([+3])
Minor Stars: horse
...
horoscope_to_prompt does the same for a horoscope at a given date.
Why not just ask the LLM to cast the chart?
Casting a chart is arithmetic, not interpretation: lunar/solar conversion, leap month handling, sexagenary cycle, the placement rules for ~100 stars, and the 四化 transformation table. A language model gets some of it right and quietly gets the rest wrong, and you cannot tell which from the output. This library does the arithmetic deterministically and verifiably, then hands the LLM the part it is actually good at — reading the chart.
Install
Rust
[dependencies]
x-iztro = "0.2"
Python — requires 3.10+, ships as an abi3 wheel with zero runtime dependencies.
pip install x-iztro
Go — the core library is embedded as WebAssembly and driven by the pure-Go wazero runtime: no cgo, no Rust toolchain, cross-compilation works as usual.
go get github.com/x-haose/x-iztro/go/iztro
Quick start
Rust
use x_iztro::{by_solar, IztroError};
use x_iztro::data::types::*;
fn main() -> Result<(), IztroError> {
let chart = by_solar(
"2000-8-16", // solar birth date
2, // hour index: 0 = early Rat hour … 12 = late Rat hour
Gender::Female,
true, // fix_leap: split leap months at the midpoint
Language::ZhCN,
Config::default(), // boundaries and school; defaults match JS iztro
)?;
// Translated strings; `soul` and `five_elements_class` are language-independent
// keys (`StarKey::PojunMaj`, `FiveElementsClass::Wood3rd`).
println!("{} / {}", chart.lunar_date, chart.chinese_date);
println!("{:?} {:?}", chart.soul, chart.five_elements_class);
// `horoscope` derefs to the data, so the six levels are plain fields.
let horoscope = chart.horoscope("2024-1-1", 0)?;
println!("{:?}", horoscope.yearly.base.mutagen);
// Bad input is an error, never a panic.
assert!(by_solar("2000-13-1", 2, Gender::Male, true, Language::ZhCN, Config::default()).is_err());
Ok(())
}
Python
from x_iztro import Astro, IztroError
from x_iztro.enums import MajorStar, Mutagen, PalaceName
chart = Astro().by_solar("2000-8-16", 2, "female")
print(chart.chinese_date, chart.soul, chart.five_elements_class)
# Enums are language-independent keys, so these checks give the same answer
# no matter which language the chart was rendered in.
soul = chart.palace(PalaceName.SOUL)
print(soul.has([MajorStar.ZIWEI]), soul.has_mutagen(Mutagen.LU))
horoscope = chart.horoscope("2024-1-1", 0)
print(horoscope.yearly.heavenly_stem, horoscope.yearly.earthly_branch)
# IztroError subclasses ValueError; .code is a machine-readable category.
try:
Astro().by_solar("2000-13-1", 2, "male")
except IztroError as e:
print(e.code) # invalid_date
Go
package main
import (
"errors"
"fmt"
"log"
"github.com/x-haose/x-iztro/go/iztro"
)
func main() {
chart, err := iztro.BySolar("2000-8-16", 2, iztro.GenderFemale, true, iztro.LanguageZhCN, nil)
if err != nil {
log.Fatal(err)
}
fmt.Println(chart.ChineseDate, chart.Soul, chart.FiveElementsClass)
soul := chart.Palace(iztro.PalaceSoul)
fmt.Println(soul.Has(iztro.StarZiweiMaj), soul.HasMutagen(iztro.MutagenLu))
horoscope, err := chart.Horoscope("2024-1-1", 0)
if err != nil {
log.Fatal(err)
}
fmt.Println(horoscope.Yearly.HeavenlyStem, horoscope.Yearly.EarthlyBranch)
// Errors carry a category you can match with errors.Is.
_, err = iztro.BySolar("2000-13-1", 2, iztro.GenderMale, true, iztro.LanguageZhCN, nil)
fmt.Println(errors.Is(err, iztro.ErrInvalidDate)) // true
}
Features
- Full chart — twelve palaces, body palace, soul/body stars, five elements class, major/minor/adjective stars with brightness and 四化 transformations.
- Six horoscope levels — decadal, yearly, monthly, daily, hourly and the childhood limit, each with its own palaces and transformations.
- Chart queries — locate a palace by name, branch or index; test stars, transformations and empty palaces; the 三方四正 surrounded-palace group; and the flying-star (飞星) family.
- Pattern judgement — 64 named star arrangements (格局), one rule set shared by natal charts and horoscope views, each hit carrying the palace it formed in, the reading that matched, and the evidencing stars. iztro has no such API.
- Knowledge packs — reading texts and school-specific star attributes live in a swappable JSON pack, not in the core. A default pack ships inside (107 stars, 64 patterns, 12 palaces, 4 transformations, 49 glossary entries, zh-CN, from iztro-docs by Sylar Long, MIT); write an overlay to replace any entry.
- Reverse lookup —
solar_dates_by_bazirecovers solar birth dates from four BaZi pillars (interpreted under the chart Config), andreverse_chartrecovers them from chart features — soul/body palace branches, five elements class, star placements, birth-year mutagens. Pruned enumeration verified by full re-charting: zero divergence from forward casting. iztro has no such API. - Two schools — the default school and 中州派 (Zhongzhou), selected per chart.
- Six languages — zh-CN, zh-TW, en-US, ja-JP, ko-KR, vi-VN, with language-independent key constants so your logic never depends on the display language.
- LLM output —
astrolabe_to_prompt/horoscope_to_promptrender a whole chart as structured text. - Validated input — date format and existence, solar years 1583–9999, hour
index 0–12. Invalid input returns
Err(IztroError)in Rust, raisesx_iztro.IztroError(aValueError) in Python, returns anerrormatchable witherrors.Isin Go, and yields{"error":"..."}JSON over the C FFI. Every failure carries a machine-readable category. Nothing panics.
Pattern judgement
chart = astro.by_solar("1985-5-3", 9, "male", language="en-US")
for hit in chart.patterns():
print(hit.name, hit.palace_name, hit.broken)
General and Wolf Together surface False
Empress and Minister Facing the Palace soul False
Marshal, Rebel and Wolf surface False
Money and Horse Galloping Together soul False
Officer and Helper Flanking Life soul False
Literary Nobility and Brilliance surface False
Literary Stars Facing Life soul True
Literary Stars in Hidden Support soul False
Literary Stars in Hidden Support soul False
The judging principles and the full table of 64 patterns are on the documentation site.
Knowledge packs
The default pack is Chinese, so pair it with a hit's language-independent key
rather than its translated name:
from x_iztro import KnowledgePack
pack = KnowledgePack.builtin()
chart = astro.by_solar("2000-8-16", 2, "female", language="en-US")
for hit in chart.patterns():
print(hit.name, "|", pack.pattern(hit.key).quotes[0])
Empress and Minister Facing the Palace | 府相朝垣命必荣
The pack format, the merge rules and how to write an overlay are on the documentation site.
Reverse lookup
from x_iztro import solar_dates_by_bazi
# every solar birth moment in 1900-2100 with the pillars 庚辰 甲申 丙午 庚寅
for c in solar_dates_by_bazi(
("gengHeavenly", "chenEarthly"), ("jiaHeavenly", "shenEarthly"),
("bingHeavenly", "wuEarthly"), ("gengHeavenly", "yinEarthly"),
):
print(c.solar_date, c.time_index)
1940-8-31 2
2000-8-16 2
2060-8-1 2
Chart-feature lookup, how the pillars follow the Config boundaries, and the truncation semantics are on the documentation site.
Accuracy
Every number is checked field-by-field against the JavaScript iztro v2.5.8 (version-pinned), with zero tolerance for differences. Roughly 710,000 golden cases in eight layers:
| Layer | Cases | Coverage |
|---|---|---|
| Tier 1 | 1,560 | 60 years × 13 hours × both genders, every field compared individually |
| Tier 2 | 37,440 | 60 years × the 1st and 15th of each month × 13 hours × both genders |
| Tier 3 | 586,430 | every day of 60 years × 13 hours × both genders × fix_leap, hashed |
| Edge years | 46,228 | the far ends of the supported range, where leap months and tables strain |
| Horoscope | 5,760 | 360 charts × 16 target dates, all six horoscope levels, every field |
| Variants | 14,268 | lunar-date charts across leap months, Zhongzhou school, all six languages |
| Config | 9,696 | each boundary switch at its non-default value |
| Astro type | 12,488 | the heaven / earth / human chart perspectives |
On top of that: the serialization contract is compared key-by-key against JS
JSON.stringify, and the three bindings are cross-checked so that the same birth
data yields the same answers in Rust, Python and Go.
cargo test # regular layers, ~15s
cargo test --release --test golden_tier3 -- --ignored # Tier 3 in full, ~20s
Configuration
Six switches, passed explicitly per chart — there is no global state.
| Switch | Values | Default | Effect |
|---|---|---|---|
year_divide |
normal / exact |
normal |
Year boundary: lunar new year, or 立春 |
horoscope_divide |
normal / exact |
normal |
Horoscope boundary: 1st of the month, or solar term |
age_divide |
normal / birthday |
normal |
Nominal age: increments at new year, or on the birthday |
day_divide |
forward / current |
forward |
Late Rat hour belongs to the next day, or the current one |
algorithm |
default / zhongzhou |
default |
School of placement rules |
astro_type |
heaven / earth / human |
heaven |
Chart perspective (Zhongzhou) |
Custom 四化 and brightness tables can be supplied alongside them.
Documentation
https://ziwei.x-hoase.com — the documentation site, in Chinese and English: a guide
that starts from zero, the Zi Wei concepts behind the data model, and per-language API
references where every function, type and method has its own entry with real output and
edge-case notes. LLM-friendly endpoints: /llms.txt,
/llms-full.txt, and any page with .md appended.
The site lives in docs/ (cd docs && npm ci && npm run dev to run it locally).
Rust API docs are also published at docs.rs/x-iztro.
Runnable projects for all three languages are under examples/.
Building from source
Only needed when changing the Rust core.
cargo build --release
# Python bindings
PYO3_USE_ABI3_FORWARD_COMPATIBILITY=1 maturin develop --features python
# Go bindings: rebuild and refresh the embedded wasm
cargo build --release --target wasm32-wasip1
cp target/wasm32-wasip1/release/x_iztro.wasm go/iztro/
Golden test data is generated from the JS iztro package:
cd tests/golden && npm install && node generate_tier1.mjs # …and the other generators
Credits
Ported from iztro by SylarLong. New to Zi Wei Dou Shu? Its author maintains an introduction at iztro.com.
License
MIT
Metadata
Release files for x-iztro 0.3.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| x_iztro-0.3.0.tar.gz | 327.7 kB | Details |
Built distributions (wheels)
| File | Reset | |||
|---|---|---|---|---|
| x_iztro-0.3.0-cp310-abi3-win_amd64.whl | CPython 3.10 | abi3 | Windows x86-64 | Details |
| x_iztro-0.3.0-cp310-abi3-manylinux_2_17_x86_64.manylinux2014_x86_64.whl | CPython 3.10 | abi3 | Linux glibc 2.17+ x86-64 | Details |
| x_iztro-0.3.0-cp310-abi3-manylinux_2_17_aarch64.manylinux2014_aarch64.whl | CPython 3.10 | abi3 | Linux glibc 2.17+ ARM64 | Details |
| x_iztro-0.3.0-cp310-abi3-macosx_10_12_x86_64.macosx_11_0_arm64.macosx_10_12_universal2.whl | CPython 3.10 | abi3 | macOS 10.12+ universal2 (ARM64, x86-64), macOS 10.12+ x86-64, macOS 11.0+ ARM64 | Details |
Total release size: 5.1 MB
Release files / x_iztro-0.3.0.tar.gz
| Download URL | x_iztro-0.3.0.tar.gz |
|---|---|
| Size | 327.7 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
4af458b2998197fba98f32f5eadeaeba3d613f21c8186861b8d73335aa090f22
|
|
BLAKE2b-256 checksum How to use checksums |
1f00bb5ab77ab6a1bfb1aaeb67e89720f0005849327926a143816c6fda22c223
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Aug 20, 2026.
Transparency logRelease files / x_iztro-0.3.0-cp310-abi3-win_amd64.whl
| Download URL | x_iztro-0.3.0-cp310-abi3-win_amd64.whl |
|---|---|
| Size | 847.5 kB |
| Tags | CPython 3.10 Windows x86-64 abi3 |
|
SHA-256 checksum How to use checksums |
a04e97ab39dcb0e616a92cd682cc00d63cee0980430239c1348efcb97fd43c6e
|
|
BLAKE2b-256 checksum How to use checksums |
49c9ef5023f35c9edc2f38170a3bc4c76a2af314c4502594b144b40057faac72
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Aug 20, 2026.
Transparency logRelease files / x_iztro-0.3.0-cp310-abi3-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
| Download URL | x_iztro-0.3.0-cp310-abi3-manylinux_2_17_x86_64.manylinux2014_x86_64.whl |
|---|---|
| Size | 1.1 MB |
| Tags | CPython 3.10 Linux glibc 2.17+ x86-64 abi3 |
|
SHA-256 checksum How to use checksums |
e12ba6f1491f0614088dec8b1545649d0e40288e184f8ee9398552ac89ae0d42
|
|
BLAKE2b-256 checksum How to use checksums |
1cce6da155e162da393749b3212a30757e1f5d5e733a0b2e36f78f5b12d9cfc1
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Aug 20, 2026.
Transparency logRelease files / x_iztro-0.3.0-cp310-abi3-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
| Download URL | x_iztro-0.3.0-cp310-abi3-manylinux_2_17_aarch64.manylinux2014_aarch64.whl |
|---|---|
| Size | 1.0 MB |
| Tags | CPython 3.10 Linux glibc 2.17+ ARM64 abi3 |
|
SHA-256 checksum How to use checksums |
14253b55ee28c4c60816f3a9c05fb9853b289a717a7ed5ca8575c45ea316638b
|
|
BLAKE2b-256 checksum How to use checksums |
1d21c09174c95467400a999fdca87921a78c44176513410109509214e56e8889
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Aug 20, 2026.
Transparency logRelease files / x_iztro-0.3.0-cp310-abi3-macosx_10_12_x86_64.macosx_11_0_arm64.macosx_10_12_universal2.whl
| Download URL | x_iztro-0.3.0-cp310-abi3-macosx_10_12_x86_64.macosx_11_0_arm64.macosx_10_12_universal2.whl |
|---|---|
| Size | 1.8 MB |
| Tags | CPython 3.10 abi3 macOS 10.12+ universal2 (ARM64, x86-64) macOS 10.12+ x86-64 macOS 11.0+ ARM64 |
|
SHA-256 checksum How to use checksums |
aa1941b7f2ae8baca4668c66bb86461a3ab02f4a22f2884cf280fd99b68f1505
|
|
BLAKE2b-256 checksum How to use checksums |
62a0242d40ff8dfe18a70b9860e2b456582e0c5d3515e45c00b77580faa614f2
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Aug 20, 2026.
Transparency log