eptr2

eptr2 (EPIAS Transparency 2.0) is a Python client for the EPIAS Transparency Platform v2.0 API, covering 232 services of the Turkish electricity and natural gas markets — prices (PTF/SMF), consumption, generation, production plans, market operations and imbalance costs.

It is an unofficial package by Robokami / Tideseed, Apache License 2.0.

📖 Full documentation: https://tideseed.github.io/eptr2/ · 🇹🇷 Türkçe bölüm aşağıda

---

Getting Started

1. Get credentials. Register with the EPIAS Transparency Platform. Your username is the registration e-mail. The platform has an English version too.

2. Install.

pip install "eptr2[allextras]"

[allextras] adds pandas (DataFrame output) and MCP server support. For a thin client with almost no dependencies, pip install eptr2 also works. Everything works with uv pip install as well.

3. Store your credentials in a .env file next to your script:

EPTR_USERNAME=your.email@example.com
EPTR_PASSWORD=yourpassword

4. Make your first call.

from eptr2 import EPTR2

eptr = EPTR2(use_dotenv=True, recycle_tgt=True)

# Market Clearing Price (MCP / PTF) for a single day
df = eptr.call("mcp", start_date="2026-07-01", end_date="2026-07-01")
print(df)

recycle_tgt=True caches the authentication ticket, so repeated runs do not log in again.

You can also pass credentials directly — handy for notebooks:

eptr = EPTR2(username="your.email@example.com", password="yourpassword")

5. Or skip Python entirely and use the command line:

eptr2 call mcp --start-date 2026-07-01 --end-date 2026-07-01 --format csv

Usage

One call pattern for all 232 services

Every service is reached with the same call method and an endpoint key:

eptr.call("mcp",         start_date="2026-07-01", end_date="2026-07-01")  # PTF
eptr.call("smp",         start_date="2026-07-01", end_date="2026-07-01")  # SMF
eptr.call("rt-cons",     start_date="2026-07-01", end_date="2026-07-01")  # Real-time consumption
eptr.call("rt-gen",      start_date="2026-07-01", end_date="2026-07-01")  # Generation by source
eptr.call("mcp-smp-imb", start_date="2026-07-01", end_date="2026-07-01")  # Imbalance prices

Turkish aliases work too: ptf → mcp, smf → smp.

Finding the right service

You do not need credentials to explore the API:

eptr2 list                     # all 232 keys grouped by category
eptr2 search dengesizlik       # keyword search, Turkish or English
eptr2 describe mcp             # parameters, method, path, description

from eptr2.agentic import search_calls, describe_call

describe_call("uevm")          # required/optional parameters and meaning

Typed convenience wrappers

If you prefer autocomplete and type hints over string keys:

from eptr2.calls import get_mcp, get_smp, get_rt_gen

df = get_mcp(start_date="2026-07-01", end_date="2026-07-01", eptr=eptr)

Composite functions

Ready-made combinations of several calls, cleaned and merged:

from eptr2.composite import (
    get_hourly_price_and_cost_data,        # MCP + SMP + WAP + imbalance/KUPST costs
    get_hourly_consumption_and_forecast_data,
    get_dabi_idm_data,                     # day-ahead + bilateral + intraday volumes
    get_bpm_range,                         # balancing market (YAL/YAT) with SMP
    get_kgup_bulk_range,                   # bulk production PLANS (by UEVCB id)
    get_rt_gen_bulk_range,                 # bulk REALIZED generation (by powerplant id)
)

df = get_hourly_price_and_cost_data(
    start_date="2026-07-01", end_date="2026-07-31", eptr=eptr
)

Note the argument order: dates first, the client as the eptr keyword.

Imbalance and KÜPST cost calculations

Pure functions — no API call, no credentials. The regulation period (pre-2026 / 2026) and the price floor/ceiling are derived from the contract code:

from eptr2.util.costs import calculate_unit_price_and_costs_by_contract

calculate_unit_price_and_costs_by_contract(
    contract="PH26070101",          # PHYYMMDDhh
    mcp=4000, smp=4000,
    system_direction="Enerji Açığı", # required when MCP == SMP
)

⚠️ When MCP equals SMP the system direction cannot be inferred from prices, and assuming a balanced system understates the negative imbalance price. Pass system_direction (or sd_sign) explicitly; the systemStatus field of mcp-smp-imb gives it.

Options for unattended / production use

eptr = EPTR2(
    use_dotenv=True,
    recycle_tgt=True,
    strict_params=True,     # raise on unknown parameters instead of warning
    connect_timeout=10.0,   # seconds (per operation, not a total deadline)
    read_timeout=60.0,
)

strict_params matters more than it looks: an unknown parameter is dropped from the request, so a misspelled filter (ppID instead of pp_id) silently turns a filtered query into an unfiltered one. By default eptr2 warns and names the correct parameter; strict_params=True makes it an error.

---

Using eptr2 with AI Agents

eptr2 ships provider-agnostic tooling — nothing is tied to one model vendor. There are three ways to use it, depending on what your agent is.

1. Chat assistants — connect the MCP server

Gives any Model Context Protocol client (VS Code agent mode, Claude Desktop/Code, Cursor, …) 18 tools: prices, consumption, generation, market operations, credential-free discovery, and pure imbalance/KÜPST calculators.

pip install "eptr2[allextras]"
eptr2 mcp-config --client vscode     # or claude-desktop, claude-code, cursor, generic

Paste the printed snippet into your client's MCP configuration, fill in your credentials, and ask questions in plain language:

"What was the market clearing price in Turkey yesterday, and how did it compare with SMP?"

2. Coding agents — let them write eptr2 code for you

Point the agent at AGENTS.md, the canonical agent guide (VS Code and Claude Code read it automatically inside this repo). Away from the repo, the agent can learn the whole API on its own, without credentials:

eptr2 schema --stdout        # machine-readable description of all 232 endpoints
eptr2 describe rt-gen        # exact parameters for one endpoint
eptr2 search üretim          # find endpoints by keyword

Install the 7 bundled skills (prices, consumption, generation, imbalance costs, market operations, API discovery, typed wrappers) into any SKILL.md-compatible runtime:

eptr2 install-skills                 # into ./.agents/skills
eptr2 install-skills --dest user     # into ~/.agents/skills
eptr2 install-skills --client claude  # into ./.claude/skills
eptr2 install-skills --client claude --dest user  # into ~/.claude/skills

The skills and MCP server are also packaged together as a portable Agent Plugin: eptr2 plugin-path.

3. Shell-driven agents — use the CLI

Data goes to stdout (JSON or CSV), diagnostics to stderr, nonzero exit codes on failure, so output pipes cleanly:

eptr2 call mcp --start-date 2026-07-01 --end-date 2026-07-01 --format json | jq '.[0]'

📚 Details: MCP server · MCP client setup · agent skills · Agent Plugin · CLI

---

Going Further

Topic	Documentation
All 232 API calls, categories and parameters	Available API Calls
Typed get_* wrapper functions (eptr2.calls)	Convenience Wrappers
Aliases, bulk calls, DataFrames	Basic Usage · DataFrames
Composite functions	Composite Functions
Imbalance / KÜPST cost calculations	Utilities API
Turkish market abbreviations (PTF, SMF, KGÜP, …)	Abbreviations
Release history	Changelog

Development versions

pip install --pre "eptr2[allextras]"                                          # latest pre-release
pip install "eptr2[allextras] @ git+https://github.com/Tideseed/eptr2.git"    # straight from GitHub

Pin a released version with eptr2==1.3.8 if you need to stay off dev builds.

---

Türkçe

eptr2 (EPİAŞ Transparency 2.0), EPİAŞ Şeffaflık Platformu 2.0 API'si üzerine geliştirilmiş bir Python paketidir. Türkiye elektrik ve doğal gaz piyasalarına ait 232 veri servisine tek bir yapıyla erişim sağlar: PTF/SMF fiyatları, tüketim, üretim, üretim planları (KGÜP/KUDÜP), piyasa işlemleri (GÖP, GİP, DGP, İA) ve dengesizlik maliyetleri.

Robokami / Tideseed tarafından geliştirilen resmi olmayan bir pakettir. Apache License 2.0 ile lisanslanmıştır; ticari kullanım dahil geniş ölçüde serbesttir.

Kurulum ve ilk adımlar

1. Kayıt. EPİAŞ Şeffaflık Platformu'na kayıt olun. Kullanıcı adınız kayıt e-postanızdır.

2. Kurulum.

pip install "eptr2[allextras]"

[allextras] seçeneği pandas (DataFrame çıktısı) ve MCP sunucu desteğini de kurar. Sadece temel istemci için pip install eptr2 yeterlidir.

3. Kimlik bilgileri. Betiğinizin yanına bir .env dosyası oluşturun:

EPTR_USERNAME=eposta@ornek.com
EPTR_PASSWORD=sifreniz

4. İlk çağrı.

from eptr2 import EPTR2

eptr = EPTR2(use_dotenv=True, recycle_tgt=True)

# Piyasa Takas Fiyatı (PTF)
df = eptr.call("ptf", start_date="2026-07-01", end_date="2026-07-01")
print(df)

recycle_tgt=True giriş biletini (TGT) saklar; her çalıştırmada yeniden giriş yapılmaz.

Kullanım

Tüm servisler aynı desenle çağrılır. Türkçe kısaltmalar takma ad olarak kullanılabilir (ptf → mcp, smf → smp):

eptr.call("ptf",         start_date="2026-07-01", end_date="2026-07-01")  # Piyasa Takas Fiyatı
eptr.call("smf",         start_date="2026-07-01", end_date="2026-07-01")  # Sistem Marjinal Fiyatı
eptr.call("rt-cons",     start_date="2026-07-01", end_date="2026-07-01")  # Gerçek zamanlı tüketim
eptr.call("rt-gen",      start_date="2026-07-01", end_date="2026-07-01")  # Kaynak bazlı üretim
eptr.call("mcp-smp-imb", start_date="2026-07-01", end_date="2026-07-01")  # Dengesizlik fiyatları

Hangi servisin ne olduğunu kimlik bilgisi gerekmeden keşfedebilirsiniz:

eptr2 list                 # 232 servisin tamamı, kategorilere göre
eptr2 search dengesizlik   # Türkçe veya İngilizce anahtar kelime araması
eptr2 describe uevm        # parametreler, yöntem, açıklama

Birden fazla çağrıyı birleştiren hazır fonksiyonlar (composite) da mevcuttur:

from eptr2.composite import get_hourly_price_and_cost_data

# PTF, SMF, AOF ve dengesizlik/KÜPST maliyetleri tek tabloda
df = get_hourly_price_and_cost_data(
    start_date="2026-07-01", end_date="2026-07-31", eptr=eptr
)

Dengesizlik ve KÜPST maliyetleri

API çağrısı gerektirmeyen saf hesaplama fonksiyonlarıdır. Mevzuat dönemi (2026 öncesi / 2026) ve taban-tavan fiyatlar sözleşme kodundan otomatik belirlenir:

from eptr2.util.costs import calculate_unit_price_and_costs_by_contract

calculate_unit_price_and_costs_by_contract(
    contract="PH26070101",           # PHYYAAGGss formatında saatlik sözleşme
    mcp=4000, smp=4000,
    system_direction="Enerji Açığı",  # PTF == SMF ise zorunlu
)

⚠️ Önemli: PTF ile SMF eşit olduğunda sistem yönü fiyatlardan anlaşılamaz. Bu durumda sistemin dengede olduğu varsayılırsa negatif dengesizlik fiyatı olduğundan düşük hesaplanır. system_direction değerini (Enerji Açığı / Enerji Fazlası / Dengede) mutlaka belirtin; bu bilgi mcp-smp-imb çağrısındaki systemStatus alanından alınabilir.

Yapay zeka ajanlarıyla kullanım

eptr2, herhangi bir yapay zeka sağlayıcısına bağlı olmayan araçlar sunar:

• MCP sunucusu — VS Code, Claude, Cursor gibi istemcilere 18 araç kazandırır. Yapılandırma çıktısı için: eptr2 mcp-config --client vscode
• Ajan becerileri (skills) — fiyat analizi, tüketim, üretim, dengesizlik maliyetleri, piyasa işlemleri ve API keşfi için 7 hazır beceri: eptr2 install-skills
• AGENTS.md — kod yazan ajanlar için başvuru rehberi
• Makine tarafından okunabilir şema — 232 servisin tamamı parametreleriyle: eptr2 schema --stdout

Böylece bir yapay zeka asistanına doğrudan "geçen ayın PTF ortalamasını hesapla" ya da "rüzgar santralim için dengesizlik maliyetini çıkar" diyebilir, ajanın sizin için doğru çağrıları yazmasını sağlayabilirsiniz.

Dokümantasyon

Ayrıntılı ve güncel dokümantasyon: https://tideseed.github.io/eptr2/