* chore: upgrade pdmt5 to v1.0.4
pdmt5 1.0.4 removes Mt5TradingClient/Mt5TradingError entirely and wraps
Mt5Config.password in pydantic SecretStr. Drop the now-dead conditional
Mt5TradingError handling in mt5cli.exceptions/sdk/retry (mt5cli already
type-checks trading clients against its own protocol, so no functional
change), and unwrap SecretStr when forwarding a base config's password to
per-account configs. Update tests and docs accordingly.
* chore: declare pydantic as a direct runtime dependency
mt5cli.sdk imports SecretStr directly from pydantic, so pin it explicitly
instead of relying on pdmt5's transitive dependency.
---------
Co-authored-by: Claude <noreply@anthropic.com>
* refactor: collapse repeated tests with pytest.mark.parametrize
Collapse 13 near-identical test methods into 4 parametrized tests
across test_cli.py and test_sdk.py, keeping all 1045 cases passing.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* feat: add fetch_recent_history_deals_for_trading_client to stable SDK
Adds a generic history deal retrieval helper for active trading clients,
a _HistoryDealsClientProtocol describing the minimal required interface,
clarified create_trading_client() docs (returns pdmt5.Mt5DataClient, not
MT5Client), 9 unit tests at 100% coverage, and updated trading.md and
public-contract.md with examples and out-of-scope strategy semantics note.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix: narrow Mt5CliClient protocol claim and preserve empty deal DataFrame schema
- _HistoryDealsClientProtocol docstring and fetch_recent_history_deals_for_trading_client
docstring now explicitly state that Mt5CliClient (mt5_session) exposes
history_deals() not history_deals_get_as_df() and does not satisfy the protocol;
the function is for trading-client sessions (pdmt5.Mt5DataClient) only
- Empty DataFrames with columns are now passed through with reset_index rather
than replaced by a bare pd.DataFrame(), preserving schema for callers that rely
on stable column names even in no-deal windows
- Tests updated to assert schema preservation on empty results and bare empty on None
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix: add combined protocol so create_trading_client() is type-safe with history deals helper
Adds _TradingHistoryDealsClientProtocol combining _Mt5ClientProtocol and
_HistoryDealsClientProtocol, and updates create_trading_client() and
mt5_trading_session() to return/yield this combined type so the natural SDK
flow `client = create_trading_client(...); fetch_recent_history_deals_for_trading_client(client)`
is type-safe under pyright strict without casts.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix: validate hours is finite before timedelta in fetch_recent_history_deals_for_trading_client
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* Bump version to 1.1.1
---------
Co-authored-by: agent <agent@localhost>
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
* feat: add Grafana-ready SQLite observability (#79, #80, #81)
New `mt5cli/grafana.py` module with idempotent DDL helpers:
- `create_snapshot_tables` — five SQLite tables for time-series account,
position, order, terminal, and run-status snapshots
- `create_grafana_views` — 13 `grafana_*` views with integer epoch-second
`time` columns; missing source tables emit warnings and are skipped
- `create_grafana_indexes` — 9 performance indexes guarded by column checks
- `ensure_grafana_schema` — convenience wrapper calling all three above
- Insert helpers: `insert_account_snapshot`, `insert_position_snapshots`,
`insert_order_snapshots`, `insert_terminal_snapshot`, `record_snapshot_run`
New stable SDK exports in `mt5cli.__init__` and `mt5cli.contract`:
- `update_observability` — appends a timestamped snapshot to a SQLite db
using an already-connected `Mt5DataClient`; never places orders
- `update_observability_with_config` — standalone wrapper that opens and
closes the MT5 connection automatically
New CLI commands (Collection panel):
- `grafana-schema` — idempotent schema setup, no MT5 connection required
- `snapshot` — append account/position/order/terminal rows; supports
`--symbol`, `--with-account/--no-account`, and equivalent flags
All public modules maintain 100 % branch coverage; 968 tests pass.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* chore: apply mdformat to docs after grafana observability additions
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* chore: bump version to 1.1.0
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix: address PR #86 review feedback
- Fix unaggregated time in grafana_realized_pnl GROUP BY query (MAX)
- Replace O(N) per-symbol API calls with single call + client-side filter
- Eliminate double create_snapshot_tables when with_grafana_schema=True
- Move grafana imports to module level in sdk.py (remove PLC0415 noqa)
- Default with_grafana_schema to False (run grafana-schema once for setup)
- Fix README position example to use snapshot_runs for latest snapshot
- Update tests to reflect new behavior and correct patch targets
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* chore: apply ruff formatting and sync lock file
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix: address owner review feedback on PR #86
- Filter grafana_*_snapshots views to only expose rows from successful
runs (JOIN snapshot_runs WHERE status='ok'), closing the partial-snapshot
visibility gap raised in PRRT_kwDORzI_286MvDJ6
- Add issubset column guards for snapshot table index creation, consistent
with the rest of create_grafana_indexes (PRRT_kwDORzI_286MvUx1)
- Fix README example queries: views expose 'time' not 'observed_at'
(PRRT_kwDORzI_286MvUxw)
- Correct public-contract.md default for with_grafana_schema (False, not
True) and point to grafana-schema command (PRRT_kwDORzI_286MvUxy)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix: add run_id to snapshot schema and fix stale-positions README query
Replace second-level observed_at as the join key between snapshot_runs
and snapshot tables with a stable run_id INTEGER PRIMARY KEY. Two runs
in the same second now get distinct run_ids, preventing view duplication
and cross-contamination from a failed run. Update README example to use
snapshot_runs for latest-snapshot lookup so zero-position runs return an
empty result instead of stale rows.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* chore: format markdown tables
Align table column widths in README and public-contract documentation.
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
* fix: close sqlite connections deterministically
* fix: require entry filter in grafana_realized_pnl and add time to grafana_trade_stats
grafana_realized_pnl now requires the entry column and filters to
close-side deals (entry IN (1, 2, 3)), consistent with grafana_symbol_pnl
and grafana_trade_stats. grafana_trade_stats now requires the time column
and emits MAX(time_expr) AS "time" so it satisfies the documented Grafana
view contract (integer epoch-second time column throughout).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix: normalize pd.Timestamp time_setup to epoch int in insert_order_snapshots
orders_get_as_df() returns datetime-converted columns by default, so
time_setup is a pd.Timestamp in normal use. Passing it directly to
sqlite3.executemany raises ProgrammingError. Added _to_epoch_int helper
that converts datetime.datetime subclasses (including pd.Timestamp) and
raw int/float values to integer epoch seconds, returning None for other
types. Regression tests cover all four input paths.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix: pre-drop all grafana_* views at start of create_grafana_views
Previously, a builder that skipped due to a missing source table or column
would not drop the view it owned, leaving stale views referencing gone
tables. Now create_grafana_views drops all 13 known grafana_* views before
calling any builder, so a schema refresh always removes views whose source
has disappeared. Regression test covers the create → drop-source → refresh
cycle.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix: expose run_id in snapshot views and drop time from summary views
Grafana snapshot views now expose run_id so latest-state queries can use
MAX(run_id) instead of the ambiguous second-level MAX(observed_at).
grafana_realized_pnl and grafana_trade_stats lose their MAX(time) column
and are reclassified as static summary views; their all-time aggregates
are not filterable by Grafana time-range selectors.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix: guard snapshot views against missing run_id and fix README view docs
_build_snapshot_view now skips with a warning when the underlying
snapshot table exists but lacks a run_id column, preventing a broken
view that fails at query time. Adds a regression test for that path.
README Grafana section now qualifies that grafana_realized_pnl and
grafana_trade_stats are static summary views (no time column) and splits
the view table to match docs/api/public-contract.md.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
---------
Co-authored-by: agent <agent@localhost>
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix: decouple mt5cli from pdmt5 high-level trading helpers
- Replace Mt5TradingClient type annotations with internal _Mt5ClientProtocol
- Lazy-import Mt5TradingClient in create_trading_client to avoid hard dependency
- Replace Mt5TradingError with Mt5OperationError in mt5cli validation paths
- Update exception handling to support future pdmt5 versions without Mt5TradingError
- Add test to enforce that mt5cli doesn't import high-level symbols at module level
- Update documentation to clarify dependency boundaries
mt5cli now relies only on low-level MT5 primitives:
- Mt5Config for configuration
- Mt5RuntimeError for runtime errors
- Raw MT5 methods (order_send, order_check, account_info, etc.)
This aligns with pdmt5's direction to remove high-level trading helpers and focus
on low-level MT5 access plus DataFrame/dict conversion.
Fixes#75 (dceoy/mt5cli#75)
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PcGVFTVgyqzse3LLw38ber
* fix: address PR #76 review feedback on pdmt5 decoupling
- Replace Mt5TradingClient with Mt5DataClient in create_trading_client()
so the function no longer depends on the high-level trading client
- Fix _RECOVERABLE_MT5_ERRORS in exceptions.py to use tuple unpacking
form, removing the incorrect ternary assignment
- Add pragma: no cover to except ImportError branches in exceptions.py
and sdk.py (dead code when pdmt5 is installed)
- Switch coverage exclude_lines to exclude_also so the default
pragma: no cover pattern is preserved; also exclude bare ... stubs
(Protocol method bodies) from coverage
- Correct inaccurate note in docs/api/public-contract.md: Mt5TradingClient
is no longer required internally; Mt5TradingError is conditionally
available but mt5cli raises Mt5OperationError for trading failures
- Update all mock patches from pdmt5.Mt5TradingClient to
mt5cli.trading.Mt5DataClient to match the new module-level import
---------
Co-authored-by: Claude <noreply@anthropic.com>
* feat: add close-positions CLI command and replace_symbol projection mode (#65#66)
Part 1 — close-positions CLI (#65):
- Add `close-positions` subcommand delegating to `close_open_positions()`.
- Accepts repeated `--symbol` and `--ticket` filters (AND semantics).
- Supports `--dry-run` (no `--yes` required); live execution requires `--yes`.
- Fails closed with `BadParameter` when neither `--symbol` nor `--ticket` is given.
- Exports normalized `OrderExecutionResult` list as a DataFrame (request/response
serialized as JSON strings for clean CSV/JSON/Parquet/SQLite output).
- `order-send` remains the raw expert path; `close-positions` is the safer
high-level helper that builds correct close requests automatically.
Part 2 — ProjectionMode and replace_symbol (#66):
- Add `ProjectionMode = Literal["add", "replace_symbol"]` type alias.
- Add optional `projection_mode` parameter to `calculate_symbol_group_margin_ratio`.
Default `"add"` preserves existing additive behavior.
`"replace_symbol"` subtracts current margin for `new_symbol`, then adds
candidate margin — the subtraction and addition are atomic (suppressed together).
- Export `ProjectionMode` from `mt5cli` and add to `STABLE_SDK_EXPORTS`.
- No mteor-specific strategy, risk-threshold, or policy logic added.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* chore: remove unused ProjectionMode import in test_contracts.py
The parametrized test_stable_exports_are_importable_from_package_root
already covers ProjectionMode via hasattr(mt5cli, name). Ruff correctly
flagged the explicit top-level import as unused (F401).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* Bump version to v0.9.6
* fix: address PR #67 review feedback
- Floor replace_symbol margin subtraction at zero to prevent negative ratio
- Serialize response unconditionally via json.dumps (null for dry-run rows)
- Return a schema-preserving empty DataFrame when results list is empty
- Add test: --dry-run --yes precedence (dry-run wins, no order_send)
- Add test: zero-match filter produces empty JSON array with exit 0
- Move projection_mode prose to stable trading section in docs
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* feat: add runtime validation for projection_mode in calculate_symbol_group_margin_ratio
Unsupported values previously silently fell through as "add". The new
_validate_projection_mode helper raises ValueError with a message that
names the bad value and the two accepted modes.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
---------
Co-authored-by: agent <agent@localhost>
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
* feat: support string login in build_config and add substitute_mapping_values (#61, #62)
Extend build_config() to accept login: int | str | None. String logins
are coerced via the existing coerce_login() helper (empty/whitespace →
None, numeric strings → int, non-numeric → ValueError). When
allow_whole_dollar_env=True, ${ENV} and $ENV placeholders are expanded
before coercion, consistent with path/password/server behavior.
Add substitute_mapping_values(), a generic recursive helper that
substitutes environment placeholders in nested dicts/lists only for
caller-selected mapping keys. Non-selected fields (including literal
dollar signs) are preserved exactly. Supports blank_string_keys_as_none
to normalise empty strings to None after substitution. No application-
specific key names (e.g. mt5_login) are hard-coded in mt5cli.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* Bump version to v0.9.5
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* docs+test: clarify substitute_mapping_values docstring and pin tuple behaviour
- Adds sentence noting list-element strings are never substituted (only
immediate dict values are), addressing reviewer finding #1.
- Rewrites Returns section to accurately describe scalar pass-through
behaviour, addressing reviewer finding #2.
- Adds recursion-depth caveat to the generic-utility docstring,
addressing reviewer finding #4.
- Adds test_tuple_container_not_traversed to pin the existing silent
tuple-exclusion contract, addressing reviewer finding #3.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* docs: update public contract and README for build_config login coercion and substitute_mapping_values
- Expands build_config row to document login: int | str | None,
numeric-string coercion, blank-string handling, and env placeholder
expansion when allow_whole_dollar_env=True.
- Adds substitute_mapping_values to the stable SDK table with a note
that key names are never hard-coded in mt5cli.
- Extends allow_whole_dollar_env paragraph to list substitute_mapping_values.
- README: adds build_config env-placeholder example and imports to the
trading lifecycle snippet.
- README: extends credential-resolution bullet with a substitute_mapping_values
usage example using generic key names.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
---------
Co-authored-by: agent <agent@localhost>
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
Add _valid_tick_price() internal helper that returns a positive finite float
from a tick dict or None for any invalid value (missing, None, NaN, infinite,
zero, negative, or unsupported type). Refactor five existing bid/ask validation
sites in trading.py to use it, removing duplicated isinstance/isfinite checks.
Add calculate_positions_margin_by_symbol() which computes margin per unique
symbol independently using the existing strict calculate_positions_margin(),
with first-seen deduplication and configurable error suppression
(Mt5TradingError, Mt5RuntimeError, AttributeError) via suppress_errors=.
Add calculate_positions_margin_safe() as a thin sum wrapper with
suppress_errors=True, returning 0.0 on empty or fully-failed inputs.
Both new helpers are exported from mt5cli, added to STABLE_SDK_EXPORTS, and
documented in docs/api/public-contract.md. Existing strict behavior of
calculate_positions_margin() is unchanged.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* feat: add fetch_latest_closed_rates_indexed and allow_whole_dollar_env opt-in (#43, #44)
Closes#43: add fetch_latest_closed_rates_indexed(client, *, symbol,
granularity, count) -> pd.DataFrame to mt5cli/trading.py. Internally
reuses fetch_latest_closed_rates_for_trading_client(), converts the
"time" column to a UTC-aware DatetimeIndex named "time", and drops the
original column. Exported from trading.__all__, mt5cli.__init__, and
STABLE_SDK_EXPORTS.
Closes#44: extend substitute_env_placeholders() with opt-in
allow_whole_dollar_env=False that expands whole-value $ENV_NAME strings
(entire string must be exactly $IDENTIFIER). Threaded through
build_config(), resolve_account_spec(), and resolve_account_specs() with
the same default=False. Partial strings like "plan$pass", "abc$ENV", or
"$ENV-suffix" are never expanded.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* chore: align Markdown table columns in docs and skill file
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix: treat numeric (float64) epoch seconds as UTC in _rate_time_to_utc
After DataFrame concat or NA upcast the time column becomes float64, which
is still epoch seconds. Using is_numeric_dtype instead of is_integer_dtype
fixes the silent misalignment. Using series.to_numpy() before passing to
pd.to_datetime avoids the redundant pd.DatetimeIndex() wrapper and aligns
with how existing rate-time normalization in schemas.py handles numeric
timestamps.
Add test_converts_float_epoch_seconds_to_utc_datetime_index to cover the
regression. Add a doc note clarifying that build_config cannot expand
login since that parameter is int | None.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* fix: reject NaT values after rate timestamp conversion in _rate_time_to_utc
pd.to_datetime() silently produces NaT for None/NaN inputs rather than
raising, so the function could return a DatetimeIndex containing NaT
despite documenting invalid timestamps as a ValueError. Check any(idx.isna())
after conversion and raise with a clear message.
Add test_raises_on_nat_time_column to cover the regression.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
* Bump version to v0.9.0
* fix: handle object numeric rate timestamps
---------
Co-authored-by: Claude <noreply@anthropic.com>
* feat: add stable SDK helpers for volume, margin, and closed bars (#39, #40, #41)
Expose generic trading utilities in the stable downstream SDK so applications
like mteor can drop local MT5 adapter code:
- normalize_order_volume() for broker step/min/max sizing
- estimate_order_margin() and calculate_positions_margin() for margin totals
- fetch_latest_closed_rates_for_trading_client() for closed bars from Mt5TradingClient
Update STABLE_SDK_EXPORTS, package-root exports, docs, and unit tests.
Co-authored-by: Daichi Narushima <dceoy@users.noreply.github.com>
* chore: bump version to 0.8.3
Co-authored-by: Daichi Narushima <dceoy@users.noreply.github.com>
* fix: address PR review feedback on volume cap, rate time, and margin grouping
- Re-apply volume_max after step normalization in normalize_order_volume()
- Drop misleading non-time index reset branch in _ensure_rate_time_column()
- Group positions by (symbol, side) before margin estimation
- Add branch-coverage tests for tick price validation and volume cap edge case
Co-authored-by: Daichi Narushima <dceoy@users.noreply.github.com>
* fix: address remaining PR review threads on docs and DatetimeIndex
- Rename unnamed DatetimeIndex column to time after reset_index()
- Guard estimate_order_margin example on positive normalized volume
- Document calculate_positions_margin skip vs error propagation behavior
- Add test for unnamed DatetimeIndex branch coverage
Co-authored-by: Daichi Narushima <dceoy@users.noreply.github.com>
* fix: harden stable SDK margin, rate fetch, and volume normalization
- Wrap order_calc_margin conversion and reject None/non-numeric results
- Validate fetched rate objects are DataFrames before time normalization
- Return 0.0 for non-finite volume inputs and constraints in normalize_order_volume
Co-authored-by: Daichi Narushima <dceoy@users.noreply.github.com>
* fix: reject non-finite volumes in margin estimation helpers
Use _is_positive_finite_number() in estimate_order_margin() and
calculate_positions_margin() so NaN/inf volumes never reach broker calls.
Add focused tests and document non-finite volume skipping in trading.md.
Co-authored-by: Daichi Narushima <dceoy@users.noreply.github.com>
* fix: guard symbol filter in calculate_positions_margin for empty frames
Return 0.0 before filtering when positions are empty or lack a symbol column.
Add regression tests for filtered calls on malformed position frames.
Co-authored-by: Daichi Narushima <dceoy@users.noreply.github.com>
---------
Co-authored-by: Cursor Agent <cursoragent@cursor.com>
Co-authored-by: Daichi Narushima <dceoy@users.noreply.github.com>
* feat: add injectable update_backend to ThrottledHistoryUpdater
Allow downstream applications to substitute the history update backend via
the ThrottledHistoryUpdater constructor without monkey-patching
mt5cli.sdk.update_history. Defaults to update_history for backward
compatibility.
Co-authored-by: Daichi Narushima <dceoy@users.noreply.github.com>
* chore: fix lint and format after QA
Co-authored-by: Daichi Narushima <dceoy@users.noreply.github.com>
* chore: bump version to 0.8.2
Co-authored-by: Daichi Narushima <dceoy@users.noreply.github.com>
* fix: use explicit None check for ThrottledHistoryUpdater backend
Only None selects the default update_history backend so falsy callable
objects with __bool__ returning False are preserved as custom backends.
Co-authored-by: Daichi Narushima <dceoy@users.noreply.github.com>
---------
Co-authored-by: Cursor Agent <cursoragent@cursor.com>
Co-authored-by: Daichi Narushima <dceoy@users.noreply.github.com>
Daichi Narushima