eviews-mcp 1.2.0__tar.gz

eviews_mcp-1.2.0/.gitignore

@@ -0,0 +1,25 @@
+__pycache__/
+*.py[cod]
+*.egg-info/
+build/
+dist/
+.venv/
+venv/
+.pytest_cache/
+.mypy_cache/
+
+# Local MCP client registration, machine specific
+.mcp.json
+
+# EViews scratch and output
+*.wf1
+*.wf2
+*.prg
+*.log
+
+# Graph and table exports produced by the examples and live tests
+*.emf
+*.wmf
+*.eps
+phillips_residuals.*
+lib_graph.*

eviews_mcp-1.2.0/CHANGELOG.md

@@ -0,0 +1,65 @@
+# Changelog
+
+## 1.2.0
+
+### Added
+- Researcher guide (`docs/EViews-Researcher-Guide.md`): a zero-to-analysis
+  walkthrough with every command and output verified against a live session.
+- `close_workfile` / `close_all_workfiles`. EViews caps how many workfiles may
+  be open and then refuses to create another.
+- `show` now reads views that freeze into a spool rather than a table, such as
+  the ARDL cointegrating relationship and error-correction results, by writing
+  the frozen object out as text.
+
+### Fixed
+- Importing a data file while a workfile was open made EViews truncate the file
+  to that page's length, silently. A 100-row file read into an open 12-row page
+  lost 88 rows with no warning. Imports now create a workfile sized to the file;
+  merging into the open page is opt-in via `into_current_page`.
+- Long title and note rows padded column 0 across the whole table, pushing the
+  numbers far to the right in unit root and diagnostic output. Such rows are now
+  written full width and excluded from column measurement.
+- `read_data` with `max_rows` below 20 sliced the head negatively and returned
+  more rows than asked for, not fewer.
+- Table cells were truncated at 40 characters with a non-ASCII ellipsis, cutting
+  text such as "Included observations: 122 after adjustments" and mangling in
+  consoles that are not UTF-8.
+
+## 1.1.0
+
+Restructured into a library with the MCP server as a thin adapter on top.
+
+### Added
+- `EViews` client class as the public Python API, usable from scripts and
+  notebooks independently of MCP. Context-manager support.
+- pandas integration: `to_dataframe` and `from_dataframe`, with the EViews page
+  frequency and span derived from a `DatetimeIndex` or `PeriodIndex`.
+- `exists`, `describe`, `table` (raw rows at full precision), `value`,
+  `obs_labels`.
+- Offline test suite that needs neither EViews nor Windows.
+- Packaging metadata, MIT licence, examples.
+
+### Fixed
+- `from_dataframe` wrote a dated frame into whatever page was already open when
+  that page was long enough, silently landing the values on the wrong dates. A
+  dated frame now governs the page.
+- Writes default to the whole page. EViews honours the active sample on writes,
+  so under a restricted sample values outside it were silently left NA.
+
+### Changed
+- Import on a non-Windows platform now raises a clear message instead of an
+  opaque `pythoncom` `ImportError`.
+
+## 1.0.0
+
+First working version, built against the EViews 13 COM type library.
+
+### Fixed relative to the original prototype
+- `GetScalar`, `PutScalar` and `GetString` do not exist on the EViews COM
+  interface and failed on every call. Replaced with `Get`, which covers all
+  three and infers the type.
+- Log capture through `output(r) "file"` never wrote a file, so every run
+  reported "Done (no output)" whether it succeeded or failed. Results are now
+  read by freezing an object into a table and pulling it across COM.
+- COM calls were made from arbitrary MCP worker threads. All EViews access is
+  now funnelled onto one apartment-initialised thread.

eviews_mcp-1.2.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Merwan Roudane
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

eviews_mcp-1.2.0/PKG-INFO

@@ -0,0 +1,300 @@
+Metadata-Version: 2.5
+Name: eviews-mcp
+Version: 1.2.0
+Summary: Drive EViews from Python, and expose it to LLM clients over MCP
+Project-URL: Homepage, https://github.com/merwanroudane/MCP_EVIEWS
+Project-URL: Repository, https://github.com/merwanroudane/MCP_EVIEWS
+Project-URL: Issues, https://github.com/merwanroudane/MCP_EVIEWS/issues
+Author-email: Merwan Roudane <merwanroudane920@gmail.com>
+Maintainer-email: Merwan Roudane <merwanroudane920@gmail.com>
+License: MIT License
+        
+        Copyright (c) 2026 Merwan Roudane
+        
+        Permission is hereby granted, free of charge, to any person obtaining a copy
+        of this software and associated documentation files (the "Software"), to deal
+        in the Software without restriction, including without limitation the rights
+        to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+        copies of the Software, and to permit persons to whom the Software is
+        furnished to do so, subject to the following conditions:
+        
+        The above copyright notice and this permission notice shall be included in all
+        copies or substantial portions of the Software.
+        
+        THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+        IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+        FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+        AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+        LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+        OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+        SOFTWARE.
+License-File: LICENSE
+Keywords: com-automation,econometrics,eviews,mcp,model-context-protocol,regression,statistics,time-series
+Classifier: Development Status :: 4 - Beta
+Classifier: Environment :: Console
+Classifier: Intended Audience :: Financial and Insurance Industry
+Classifier: Intended Audience :: Science/Research
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: Microsoft :: Windows
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Topic :: Office/Business :: Financial :: Investment
+Classifier: Topic :: Scientific/Engineering :: Information Analysis
+Requires-Python: >=3.10
+Requires-Dist: mcp[cli]>=1.2.0
+Requires-Dist: pywin32>=306; sys_platform == 'win32'
+Provides-Extra: dev
+Requires-Dist: pandas>=1.5; extra == 'dev'
+Requires-Dist: pytest>=7.0; extra == 'dev'
+Provides-Extra: pandas
+Requires-Dist: pandas>=1.5; extra == 'pandas'
+Description-Content-Type: text/markdown
+
+# eviews-mcp
+
+Drive **EViews** from Python, and expose it to LLM clients over the Model
+Context Protocol.
+
+Two things in one package:
+
+- **A library.** An `EViews` class for scripts and notebooks — build workfiles,
+  estimate models, read results back as text or pandas DataFrames.
+- **An MCP server.** The same capabilities as tools, so an assistant can do
+  econometrics in a real EViews session.
+
+Built and tested against **EViews 13** on Windows; EViews 10–14 resolve
+correctly through the same COM interface.
+
+> **New to this?** The [**EViews Researcher Guide**](https://github.com/merwanroudane/MCP_EVIEWS/blob/main/docs/EViews-Researcher-Guide.md)
+> takes you from a clean machine to a finished ARDL study, with every command and
+> every output verified against a real EViews session. No Python knowledge assumed.
+
+## Install
+
+```bash
+pip install git+https://github.com/merwanroudane/MCP_EVIEWS.git
+```
+
+With pandas support:
+
+```bash
+pip install "eviews-mcp[pandas] @ git+https://github.com/merwanroudane/MCP_EVIEWS.git"
+```
+
+Or from a clone, for development:
+
+```bash
+git clone https://github.com/merwanroudane/MCP_EVIEWS.git
+cd MCP_EVIEWS
+pip install -e .[dev]
+```
+
+Requires Windows and a local EViews installation, since it drives EViews
+through COM automation.
+
+## Library use
+
+```python
+from eviews_mcp import EViews
+
+with EViews() as ev:
+    ev.create_workfile("q", "1990q1", "2020q4")
+    ev.run("""
+        series k   = 100 + @trend + 3*@nrnd
+        series l   = 50 + 0.5*@trend + 2*@nrnd
+        series gdp = 10 + 0.6*k + 0.3*l + 2*@nrnd
+        equation eq1.ls gdp c k l
+    """)
+
+    print(ev.show("eq1"))
+    print(ev.value("eq1.@r2"))
+```
+
+```text
+Dependent Variable: GDP
+Method: Least Squares
+Included observations: 124
+
+Variable      Coefficient   Std. Error   t-Statistic   Prob.
+
+C             9.58073       0.829252     11.5535       2.90e-21
+K             0.591693      0.0333819    17.7250       1.99e-35
+L             0.322394      0.0673315     4.78816      4.81e-06
+
+R-squared     0.994702      Mean dependent var         131.080
+```
+
+### Any EViews view, as text
+
+`show` takes a view, so diagnostics need no extra API:
+
+```python
+ev.show("eq1", "wald c(2)=c(3)")   # coefficient restriction test
+ev.show("eq1", "resids(t)")        # residual table
+ev.show("eq1", "coefcov")          # coefficient covariance
+ev.show("eq1", "auto(2)")          # Breusch-Godfrey serial correlation
+ev.show("eq1", "white")            # White heteroskedasticity test
+ev.show("gdp", "uroot")            # unit root test
+ev.show("gdp", "correl")           # correlogram
+ev.show("ardl1", "cointrel")       # ARDL long-run relationship
+ev.show("var1", "impulse(t)")      # impulse response table
+ev.show("var1", "testexog")        # Granger causality
+```
+
+`resids` and `impulse` draw graphs by default; the `(t)` variants ask EViews
+for the table form. Views that freeze into a spool rather than a table -- the
+ARDL cointegrating relationship among them -- cannot be read over COM at all,
+so `show` falls back to a text dump for those.
+
+For the numbers rather than the layout, `table()` returns raw rows at full
+double precision, and `value()` returns one number:
+
+```python
+rows = ev.table("eq1")             # tuple of row tuples
+r2   = ev.value("eq1.@r2")         # 0.9947015...
+beta = ev.value("eq1.@coefs(2)")
+```
+
+### pandas both ways
+
+```python
+frame = ev.to_dataframe(["gdp", "k", "l"])   # indexed 1990Q1, 1990Q2, ...
+frame.corr()
+```
+
+Writing back, a `DatetimeIndex` or `PeriodIndex` decides the page frequency and
+span, so dates stay aligned:
+
+```python
+import pandas as pd
+
+index = pd.period_range("2005Q1", periods=40, freq="Q")
+df = pd.DataFrame({"inflation": ..., "unemployment": ...}, index=index)
+
+ev.from_dataframe(df)              # creates a quarterly 2005Q1–2014Q4 page
+ev.run("equation phillips.ls inflation c unemployment")
+```
+
+Non-numeric columns are skipped rather than failing the whole frame.
+
+### Graphs
+
+Graphs cannot be rendered as text, so write them to a file:
+
+```python
+ev.export_object("phillips", "residuals.png", view="resids")
+```
+
+Graph formats: `png`, `jpg`, `pdf`, `emf`, `wmf`, `bmp`, `gif`, `eps`, `tex`.
+Table formats: `csv`, `rtf`, `txt`, `html`.
+
+### Errors
+
+EViews reports failures precisely, including the line number inside a program,
+and those messages are passed through unchanged:
+
+```python
+ev.run("series ok = 1\nbroken_command\n")
+```
+
+```text
+EViewsError: BROKEN_COMMAND is not defined or is an illegal command
+in "BROKEN_COMMAND" in MCP_3BA1F7DC.PRG on line 2.
+```
+
+## MCP server use
+
+Register the `eviews-mcp` command with your MCP client:
+
+```json
+{
+  "mcpServers": {
+    "eviews": {
+      "command": "eviews-mcp"
+    }
+  }
+}
+```
+
+For Claude Code:
+
+```bash
+claude mcp add eviews -- eviews-mcp
+```
+
+### Tools
+
+| Tool | Purpose |
+|---|---|
+| `eviews_status` | Connection, version, active workfile. Start here when debugging. |
+| `reset_eviews` | Discard the instance and start clean. |
+| `set_eviews_visible` | Show or hide the EViews window. |
+| `create_workfile` | New page by frequency and range. |
+| `open_workfile` / `save_workfile` | Open and save `.wf1` / `.wf2`. |
+| `close_workfile` | Close one or all open workfiles. |
+| `workfile_info` | Name, page, frequency, range, sample. |
+| `list_objects` | Inventory, filterable by EViews type. |
+| `set_sample` | Set the estimation sample. |
+| `run_eviews_code` | **Main tool.** Run a block of EViews program code. |
+| `run_program_file` | Run an existing `.prg`, with arguments. |
+| `command` | A single command line. |
+| `show` | **Render any object as a text table.** |
+| `evaluate` | One value from an expression. |
+| `describe_object` | Type, plus statistics for a series. |
+| `read_data` | Series as an aligned table or full-precision CSV. |
+| `write_series` | Write values into a series. |
+| `import_data` | Read `.xlsx`, `.csv`, `.dta`, `.sav`, and more. |
+| `export_data` | Write series to a file. |
+| `export_object` | Save an object — the way to retrieve graphs. |
+
+Results are not echoed by `run_eviews_code`, because EViews sends program
+output to its own log window where COM cannot reach it. Estimate into a named
+object and call `show` on it.
+
+## Behaviour worth knowing
+
+These are EViews characteristics that the library handles for you, documented
+because they surprise people writing COM code directly.
+
+- **Writes respect the active sample.** Under `smpl 2000m3 2000m6`, writing 12
+  values lands 4 and silently leaves the rest NA. Writes therefore default to
+  the whole page; pass `sample=""` to opt into the current sample instead.
+- **`save` ignores the file extension.** `graph.save "out.png"` writes EMF
+  data. The format is passed explicitly, and a save that produces no file
+  raises rather than reporting success.
+- **Relative paths resolve against EViews**, not the calling process, so paths
+  are made absolute before they are handed over.
+- **A dated frame governs the page.** Writing a 12-row quarterly frame into an
+  open 80-row page would land the values on the wrong dates, so a page matching
+  the frame is created instead.
+- **No log redirection.** The `output` command requires a frozen object name
+  and otherwise writes nothing at all, so it cannot capture a log. Results come
+  from freezing an object into a table and reading that.
+- **No `GetScalar` / `PutScalar` / `GetString`.** These are not on the EViews
+  COM interface at all. `Get` covers them and infers the type.
+- **One COM thread.** MCP dispatches synchronous tools across a thread pool,
+  and a COM pointer is not valid across apartments, so every EViews call is
+  funnelled onto a single apartment-initialised thread.
+- **Importing into an open workfile truncates the file** to that page's length,
+  silently. Imports therefore create a new workfile by default; merging into the
+  current page is opt-in.
+- **EViews limits how many workfiles may be open** and then refuses to create
+  another, so `close_workfile` exists to keep long sessions healthy.
+
+## Tests
+
+```bash
+python tests/test_offline.py        # 20 tests, no EViews needed
+python tests/test_live.py           # 25 tests, drives the MCP tool layer
+python tests/test_live_client.py    # 39 tests, drives the library API
+```
+
+`pytest` runs the offline suite by default; the live suites are opt-in because
+they need an EViews licence.
+
+## Licence
+
+MIT. Copyright (c) 2026 Merwan Roudane.

eviews_mcp-1.2.0/README.md

@@ -0,0 +1,245 @@
+# eviews-mcp
+
+Drive **EViews** from Python, and expose it to LLM clients over the Model
+Context Protocol.
+
+Two things in one package:
+
+- **A library.** An `EViews` class for scripts and notebooks — build workfiles,
+  estimate models, read results back as text or pandas DataFrames.
+- **An MCP server.** The same capabilities as tools, so an assistant can do
+  econometrics in a real EViews session.
+
+Built and tested against **EViews 13** on Windows; EViews 10–14 resolve
+correctly through the same COM interface.
+
+> **New to this?** The [**EViews Researcher Guide**](https://github.com/merwanroudane/MCP_EVIEWS/blob/main/docs/EViews-Researcher-Guide.md)
+> takes you from a clean machine to a finished ARDL study, with every command and
+> every output verified against a real EViews session. No Python knowledge assumed.
+
+## Install
+
+```bash
+pip install git+https://github.com/merwanroudane/MCP_EVIEWS.git
+```
+
+With pandas support:
+
+```bash
+pip install "eviews-mcp[pandas] @ git+https://github.com/merwanroudane/MCP_EVIEWS.git"
+```
+
+Or from a clone, for development:
+
+```bash
+git clone https://github.com/merwanroudane/MCP_EVIEWS.git
+cd MCP_EVIEWS
+pip install -e .[dev]
+```
+
+Requires Windows and a local EViews installation, since it drives EViews
+through COM automation.
+
+## Library use
+
+```python
+from eviews_mcp import EViews
+
+with EViews() as ev:
+    ev.create_workfile("q", "1990q1", "2020q4")
+    ev.run("""
+        series k   = 100 + @trend + 3*@nrnd
+        series l   = 50 + 0.5*@trend + 2*@nrnd
+        series gdp = 10 + 0.6*k + 0.3*l + 2*@nrnd
+        equation eq1.ls gdp c k l
+    """)
+
+    print(ev.show("eq1"))
+    print(ev.value("eq1.@r2"))
+```
+
+```text
+Dependent Variable: GDP
+Method: Least Squares
+Included observations: 124
+
+Variable      Coefficient   Std. Error   t-Statistic   Prob.
+
+C             9.58073       0.829252     11.5535       2.90e-21
+K             0.591693      0.0333819    17.7250       1.99e-35
+L             0.322394      0.0673315     4.78816      4.81e-06
+
+R-squared     0.994702      Mean dependent var         131.080
+```
+
+### Any EViews view, as text
+
+`show` takes a view, so diagnostics need no extra API:
+
+```python
+ev.show("eq1", "wald c(2)=c(3)")   # coefficient restriction test
+ev.show("eq1", "resids(t)")        # residual table
+ev.show("eq1", "coefcov")          # coefficient covariance
+ev.show("eq1", "auto(2)")          # Breusch-Godfrey serial correlation
+ev.show("eq1", "white")            # White heteroskedasticity test
+ev.show("gdp", "uroot")            # unit root test
+ev.show("gdp", "correl")           # correlogram
+ev.show("ardl1", "cointrel")       # ARDL long-run relationship
+ev.show("var1", "impulse(t)")      # impulse response table
+ev.show("var1", "testexog")        # Granger causality
+```
+
+`resids` and `impulse` draw graphs by default; the `(t)` variants ask EViews
+for the table form. Views that freeze into a spool rather than a table -- the
+ARDL cointegrating relationship among them -- cannot be read over COM at all,
+so `show` falls back to a text dump for those.
+
+For the numbers rather than the layout, `table()` returns raw rows at full
+double precision, and `value()` returns one number:
+
+```python
+rows = ev.table("eq1")             # tuple of row tuples
+r2   = ev.value("eq1.@r2")         # 0.9947015...
+beta = ev.value("eq1.@coefs(2)")
+```
+
+### pandas both ways
+
+```python
+frame = ev.to_dataframe(["gdp", "k", "l"])   # indexed 1990Q1, 1990Q2, ...
+frame.corr()
+```
+
+Writing back, a `DatetimeIndex` or `PeriodIndex` decides the page frequency and
+span, so dates stay aligned:
+
+```python
+import pandas as pd
+
+index = pd.period_range("2005Q1", periods=40, freq="Q")
+df = pd.DataFrame({"inflation": ..., "unemployment": ...}, index=index)
+
+ev.from_dataframe(df)              # creates a quarterly 2005Q1–2014Q4 page
+ev.run("equation phillips.ls inflation c unemployment")
+```
+
+Non-numeric columns are skipped rather than failing the whole frame.
+
+### Graphs
+
+Graphs cannot be rendered as text, so write them to a file:
+
+```python
+ev.export_object("phillips", "residuals.png", view="resids")
+```
+
+Graph formats: `png`, `jpg`, `pdf`, `emf`, `wmf`, `bmp`, `gif`, `eps`, `tex`.
+Table formats: `csv`, `rtf`, `txt`, `html`.
+
+### Errors
+
+EViews reports failures precisely, including the line number inside a program,
+and those messages are passed through unchanged:
+
+```python
+ev.run("series ok = 1\nbroken_command\n")
+```
+
+```text
+EViewsError: BROKEN_COMMAND is not defined or is an illegal command
+in "BROKEN_COMMAND" in MCP_3BA1F7DC.PRG on line 2.
+```
+
+## MCP server use
+
+Register the `eviews-mcp` command with your MCP client:
+
+```json
+{
+  "mcpServers": {
+    "eviews": {
+      "command": "eviews-mcp"
+    }
+  }
+}
+```
+
+For Claude Code:
+
+```bash
+claude mcp add eviews -- eviews-mcp
+```
+
+### Tools
+
+| Tool | Purpose |
+|---|---|
+| `eviews_status` | Connection, version, active workfile. Start here when debugging. |
+| `reset_eviews` | Discard the instance and start clean. |
+| `set_eviews_visible` | Show or hide the EViews window. |
+| `create_workfile` | New page by frequency and range. |
+| `open_workfile` / `save_workfile` | Open and save `.wf1` / `.wf2`. |
+| `close_workfile` | Close one or all open workfiles. |
+| `workfile_info` | Name, page, frequency, range, sample. |
+| `list_objects` | Inventory, filterable by EViews type. |
+| `set_sample` | Set the estimation sample. |
+| `run_eviews_code` | **Main tool.** Run a block of EViews program code. |
+| `run_program_file` | Run an existing `.prg`, with arguments. |
+| `command` | A single command line. |
+| `show` | **Render any object as a text table.** |
+| `evaluate` | One value from an expression. |
+| `describe_object` | Type, plus statistics for a series. |
+| `read_data` | Series as an aligned table or full-precision CSV. |
+| `write_series` | Write values into a series. |
+| `import_data` | Read `.xlsx`, `.csv`, `.dta`, `.sav`, and more. |
+| `export_data` | Write series to a file. |
+| `export_object` | Save an object — the way to retrieve graphs. |
+
+Results are not echoed by `run_eviews_code`, because EViews sends program
+output to its own log window where COM cannot reach it. Estimate into a named
+object and call `show` on it.
+
+## Behaviour worth knowing
+
+These are EViews characteristics that the library handles for you, documented
+because they surprise people writing COM code directly.
+
+- **Writes respect the active sample.** Under `smpl 2000m3 2000m6`, writing 12
+  values lands 4 and silently leaves the rest NA. Writes therefore default to
+  the whole page; pass `sample=""` to opt into the current sample instead.
+- **`save` ignores the file extension.** `graph.save "out.png"` writes EMF
+  data. The format is passed explicitly, and a save that produces no file
+  raises rather than reporting success.
+- **Relative paths resolve against EViews**, not the calling process, so paths
+  are made absolute before they are handed over.
+- **A dated frame governs the page.** Writing a 12-row quarterly frame into an
+  open 80-row page would land the values on the wrong dates, so a page matching
+  the frame is created instead.
+- **No log redirection.** The `output` command requires a frozen object name
+  and otherwise writes nothing at all, so it cannot capture a log. Results come
+  from freezing an object into a table and reading that.
+- **No `GetScalar` / `PutScalar` / `GetString`.** These are not on the EViews
+  COM interface at all. `Get` covers them and infers the type.
+- **One COM thread.** MCP dispatches synchronous tools across a thread pool,
+  and a COM pointer is not valid across apartments, so every EViews call is
+  funnelled onto a single apartment-initialised thread.
+- **Importing into an open workfile truncates the file** to that page's length,
+  silently. Imports therefore create a new workfile by default; merging into the
+  current page is opt-in.
+- **EViews limits how many workfiles may be open** and then refuses to create
+  another, so `close_workfile` exists to keep long sessions healthy.
+
+## Tests
+
+```bash
+python tests/test_offline.py        # 20 tests, no EViews needed
+python tests/test_live.py           # 25 tests, drives the MCP tool layer
+python tests/test_live_client.py    # 39 tests, drives the library API
+```
+
+`pytest` runs the offline suite by default; the live suites are opt-in because
+they need an EViews licence.
+
+## Licence
+
+MIT. Copyright (c) 2026 Merwan Roudane.