notlang 0.1.0

GIS uchun soddalashtirilgan buyruqlar tili (Python ustida ishlaydi) — Uzbek-friendly domain-specific language for GIS/RS

notLang

GIS uchun soddalashtirilgan buyruqlar tili. Python ustida ishlaydi, natijada uzun GDAL/rasterio/geopandas kodi o'rniga 2-3 qatorlik notLang skripti yeziladi.

O'rnatish

Eng qulay usul — GitHub'dan to'g'ridan-to'g'ri (zip yuklash shart emas):

pip install git+https://github.com/<GITHUB_USER>/notlang.git

Colab/Jupyter'da (GEE va Jupyter qo'llab-quvvatlashi bilan):

!pip install "notlang[gee,jupyter] @ git+https://github.com/<GITHUB_USER>/notlang.git"

Mahalliy (development) rejimida:

pip install -e .

O'rnatgandan so'ng notlang komandasi terminalda, PowerShell'da va CMD'da ishlaydi:

notlang                # interaktiv rejim (REPL)
notlang skript.nl       # skript faylini ishga tushirish
notlang --version       # versiyani ko'rsatish

Holat: Faza 2 (interpretator ulandi — notLang endi ishlaydi!)

Endi notlang haqiqiy buyruqlarni bajaradi. Misol skript:

load "hudud.geojson" as a
info a
stats a
show a to "hudud.png"

load "mask.geojson" as m
clip a with m -> a_clip
save a_clip to "natija.geojson"

buffer a 500 -> a_buf
reproject a to 3857 -> a_3857

Ishga tushirish:

notlang skript.nl        # faylni ishga tushirish
notlang                  # interaktiv rejim (REPL), qator-qator yozish

REPL'da help yozsangiz barcha buyruqlar ro'yxati chiqadi.

Qo'llab-quvvatlanadigan buyruqlar (Faza 2)

Buyruq	Misol
load	load "a.shp" as x
show	show x title "Sarlavha" color "red" to "chiqish.png"
info	info x
stats	stats x
list	list x shapeName
save	save x to "chiqish.geojson"
filter	filter x where "shapeName == 'Xorazm'" -> y
clip	clip x with mask -> y
merge	merge x y -> z (ikkita vektor)
mosaic	mosaic x y -> z (ikkita rastr)
ndvi	ndvi x 3 4 -> z (red-band nir-band)
zonal_stats	zonal_stats raster with zonalar -> t (min/max/mean/std)
chart	chart x ustun bar to "grafik.png" (bar/hist/line)
install	install "paket-nomi" (kerakli kutubxonani o'rnatadi)
buffer	buffer x 500 -> y
reproject	reproject x to 4326 -> y

Sinash uchun:

python make_test_data.py        # sinov fayllarini yaratadi
python test_core.py             # runtime funksiyalarini sinaydi
notlang examples/skript.nl      # to'liq notLang skriptini ishga tushiradi

Keyingi bosqichlar rejasi uchun loyihaning to'liq hujjatiga qarang (Faza 5: GEE/Copernicus ulanish).

Jupyter / Google Colab'da ishlatish (Faza 4)

Avval o'rnating (Jupyter/Colab qo'llab-quvvatlashi bilan):

pip install -e ".[jupyter]"

Notebook katagida:

%load_ext notlang

Keyin har qanday katakda:

%%notlang
load "hudud.geojson" as a
info a
show a

O'zgaruvchilar (a kabi) keyingi kataklarda ham ishlatilaveradi — xuddi REPL'dagidek, bitta sessiya davomida saqlanib qoladi.

Bitta qatorlik buyruq uchun (katak emas, bitta qator):

%notlang_line show a

Google Colab'da ham xuddi shunday ishlaydi — avval !pip install notlang[jupyter] (yoki paketni yuklab, !pip install -e .), so'ng %load_ext notlang deb yozish kifoya.

Google Earth Engine'ga ulanish (Faza 5)

Avval o'rnating:

pip install -e ".[gee]"

Ishlatish (REPL, skript yoki %%notlang — barchasida bir xil):

connect gee

Birinchi marta ishlatganda brauzerda Google login oynasi ochiladi (Colab'da bu link ko'rinishida chiqadi — uni bosib, Google hisobingiz bilan ruxsat berasiz). Sizda Earth Engine yoqilgan Google Cloud loyihasi (project) bo'lishi kerak — https://code.earthengine.google.com/ orqali ro'yxatdan o'ting.

Tasvir olish:

load "hudud.geojson" as region
connect gee
get "COPERNICUS/S2_SR_HARMONIC" date "2024-01-01" "2024-02-01" bbox region -> img
info img
show img

get buyrug'i berilgan kolleksiyadan (dataset), sana oralig'ida, region qatlamining chegara qutisi (bounding box) ichidagi barcha tasvirlarning median kompozitini yuklab, mahalliy rastr sifatida qaytaradi. Katta hududlar yoki uzoq sana oralig'i GEE eksport chegarasiga (odatda ~10-50 MB) tegishi mumkin — shunday xato chiqsa, hududni yoki sana oralig'ini kichraytiring.

Diqqat: bu qismni men (Claude) o'z muhitimda haqiqiy Google hisobi bilan sinab ko'ra olmadim — internetga chiqish cheklangan. Kod sintaksis va mantiq jihatidan tekshirilgan (xato holatlari to'g'ri ushlanadi), lekin haqiqiy autentifikatsiya va tasvir yuklashni o'zingizning muhitingizda (Colab tavsiya etiladi) sinab ko'rishingiz kerak.

Copernicus Data Space'ga ulanish (Faza 5b)

Dataspace.copernicus.eu da ro'yxatdan o'ting, so'ng User Settings -> OAuth clients bo'limidan client_id va client_secret oling.

connect copernicus "client_id" "client_secret"

Tasvir olish (Sentinel-2 L2A, ko'k/yashil/qizil/NIR bandlari — B02, B03, B04, B08):

load "hudud.geojson" as region
connect copernicus "client_id" "client_secret"
get_sentinel date "2024-01-01" "2024-02-01" bbox region scale 100 -> img
info img
show img

Bu autentifikatsiya OAuth client_credentials oqimi orqali ishlaydi (brauzer kerak emas, GEE'dan farqli). Sinovda ulanish kodi haqiqiy Copernicus serveriga yetib bordi va noto'g'ri kredensial uchun to'g'ri 403 xatosi qaytardi — demak so'rov formati to'g'ri; siz faqat haqiqiy client_id/client_secret bilan sinashingiz kerak.