# dsolve
`dsolve` is a package to solve systems of dynamic equations in Python.
## Sequence Space
$$F(X,\mathcal{E})=0$$
$$f_t(x_{t-1},x_{t},x_{t+1},\epsilon_t)=0\qquad \forall t$$
## Symbolic
A package to solve systems of dynamic equations with Python. It understands $\LaTeX$ syntax and it requires minimum specifications from the user end. It solves problems of the form:
$$A_0\begin{bmatrix}x_{t+1}\\ E_{t}[p_{t+1}]\end{bmatrix}=A_1\begin{bmatrix}x_{t}\\ p_{t}\end{bmatrix}+\gamma z_t$$
with $x_t$ given. Following Blanchard Kahn notation, $x_{t}$ are state variables (known at time $t$) while $p_{t}$ are forward-looking variables, and $z_t$ are shocks with $E_t[z_{t+1}]=0$. The solver uses the Klein (2000) algorithm which allows for $A_0$ to be invertible.
Returns the matrix solution
$$p_t=\Theta_p x_t+Nz_t$$
$$x_{t+1}=\Theta_x x_t+Lz_t$$
and methods to plot impulse responses given a sequence of $z_t$
The main class of the package is `Klein`, which stores and solves the dynamic system. It takes a list of strings that are written as $\LaTeX$ equations, a dictionary that define the numeric values of the parameters, and the specification of `x`, `p` and `z`, specified as a list of $\LaTeX$ strings or a long string separated by commas.
Usage (for more examples check the [notebook tutorial](https://github.com/marcdelabarrera/dsolve/blob/main/notebooks/dsolve_tutorial.ipynb))
```python
from dsolve.solvers import Klein
# Your latex equations here as a list of strings
eq=[
'\pi_{t}=\beta*E\pi_{t+1}+\kappa*y_{t}+u_{t}',
'y_{t}=Ey_{t+1}+(1-\phi)*E[\pi_{t+1}]+\epsilon_{t}',
'\epsilon_{t} = \rho_v*\epsilon_{t-1}+v_{t}'
]
# Your calibration here as a dictionary
calibration = {'\beta':0.98,'\kappa':0.1,'\phi':1.1,'\rho_v':0.8}
# Define pre-determined variables, forward looking variables, and shocks as strings separated by commas or a list of strings.
x = '\epsilon_{t-1}'
p = '\pi_t, y_t'
z = 'v_t, u_t'
system = Klein(eq = eq, x=x, p=p, z=z, calibration=calibration)
# Simulate the inpulse response of a shock v_{0}=0 for 12 periods when \epsilon_{-1}=0
system.simulate(x0=0, z = {'v_{t}':1}, T=12)
```
## Flexible input reading
The standarized way to write a variable is `E_{t}[x_{s}]` to represent the expectation of `x_{s}` at time `t`. but `dsolve` understands other formats. `Ex_{s}`, `E[x_s]` and `Ex_s` are quivalents to `E_{t}[x_{s}]`, and the subscript `t` is assumed.
Greek symbols can be writen as `\rho` or just `rho`.
`dsolve` understands fractions and sums. `\sum_{i=0}^{2}{x_{i,t}}` produces `x_{0,t}+x_{1,t}+x_{2,t}` and fraction `\frac{a}{b}` produces `(a)/(b)`
Raw data
{
"_id": null,
"home_page": "https://github.com/marcdelabarrera/dsolve",
"name": "dsolve",
"maintainer": "",
"docs_url": null,
"requires_python": "",
"maintainer_email": "",
"keywords": "",
"author": "Marc de la Barrera i Bardalet",
"author_email": "mbarrera@mit.edu",
"download_url": "https://files.pythonhosted.org/packages/fa/74/5b93be8027b5a2ca6b8d259d8e2b52bd671c3a9c05787cffb4a2c38a5723/dsolve-0.0.16.tar.gz",
"platform": null,
"description": "# dsolve\n\n`dsolve` is a package to solve systems of dynamic equations in Python. \n\n## Sequence Space\n\n$$F(X,\\mathcal{E})=0$$\n\n$$f_t(x_{t-1},x_{t},x_{t+1},\\epsilon_t)=0\\qquad \\forall t$$\n\n\n\n## Symbolic\nA package to solve systems of dynamic equations with Python. It understands $\\LaTeX$ syntax and it requires minimum specifications from the user end. It solves problems of the form:\n\n$$A_0\\begin{bmatrix}x_{t+1}\\\\ E_{t}[p_{t+1}]\\end{bmatrix}=A_1\\begin{bmatrix}x_{t}\\\\ p_{t}\\end{bmatrix}+\\gamma z_t$$\n\nwith $x_t$ given. Following Blanchard Kahn notation, $x_{t}$ are state variables (known at time $t$) while $p_{t}$ are forward-looking variables, and $z_t$ are shocks with $E_t[z_{t+1}]=0$. The solver uses the Klein (2000) algorithm which allows for $A_0$ to be invertible. \n\nReturns the matrix solution\n\n\n$$p_t=\\Theta_p x_t+Nz_t$$\n$$x_{t+1}=\\Theta_x x_t+Lz_t$$\n\nand methods to plot impulse responses given a sequence of $z_t$\n\nThe main class of the package is `Klein`, which stores and solves the dynamic system. It takes a list of strings that are written as $\\LaTeX$ equations, a dictionary that define the numeric values of the parameters, and the specification of `x`, `p` and `z`, specified as a list of $\\LaTeX$ strings or a long string separated by commas. \n\nUsage (for more examples check the [notebook tutorial](https://github.com/marcdelabarrera/dsolve/blob/main/notebooks/dsolve_tutorial.ipynb))\n```python\nfrom dsolve.solvers import Klein\n\n# Your latex equations here as a list of strings\neq=[\n '\\pi_{t}=\\beta*E\\pi_{t+1}+\\kappa*y_{t}+u_{t}',\n 'y_{t}=Ey_{t+1}+(1-\\phi)*E[\\pi_{t+1}]+\\epsilon_{t}',\n '\\epsilon_{t} = \\rho_v*\\epsilon_{t-1}+v_{t}'\n]\n\n# Your calibration here as a dictionary\ncalibration = {'\\beta':0.98,'\\kappa':0.1,'\\phi':1.1,'\\rho_v':0.8}\n\n# Define pre-determined variables, forward looking variables, and shocks as strings separated by commas or a list of strings.\n\nx = '\\epsilon_{t-1}'\np = '\\pi_t, y_t'\nz = 'v_t, u_t'\n\nsystem = Klein(eq = eq, x=x, p=p, z=z, calibration=calibration)\n\n# Simulate the inpulse response of a shock v_{0}=0 for 12 periods when \\epsilon_{-1}=0\n\nsystem.simulate(x0=0, z = {'v_{t}':1}, T=12)\n```\n\n## Flexible input reading\n\nThe standarized way to write a variable is `E_{t}[x_{s}]` to represent the expectation of `x_{s}` at time `t`. but `dsolve` understands other formats. `Ex_{s}`, `E[x_s]` and `Ex_s` are quivalents to `E_{t}[x_{s}]`, and the subscript `t` is assumed. \n\nGreek symbols can be writen as `\\rho` or just `rho`. \n\n`dsolve` understands fractions and sums. `\\sum_{i=0}^{2}{x_{i,t}}` produces `x_{0,t}+x_{1,t}+x_{2,t}` and fraction `\\frac{a}{b}` produces `(a)/(b)`\n",
"bugtrack_url": null,
"license": "",
"summary": "Solver of dynamic equations with forward looking variables",
"version": "0.0.16",
"split_keywords": [],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "1d4f8987486cc7f4581b54f58595f93a3039ddf76dfa48e097bca89764ef4462",
"md5": "e575cadc79b3f3f86f35489119fbb4e2",
"sha256": "48d840e1fd88b6167d2341c8e0395c89ea28fca5a4c3d7b523495b925c01d114"
},
"downloads": -1,
"filename": "dsolve-0.0.16-py3-none-any.whl",
"has_sig": false,
"md5_digest": "e575cadc79b3f3f86f35489119fbb4e2",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": null,
"size": 16485,
"upload_time": "2023-04-24T14:32:09",
"upload_time_iso_8601": "2023-04-24T14:32:09.618046Z",
"url": "https://files.pythonhosted.org/packages/1d/4f/8987486cc7f4581b54f58595f93a3039ddf76dfa48e097bca89764ef4462/dsolve-0.0.16-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "fa745b93be8027b5a2ca6b8d259d8e2b52bd671c3a9c05787cffb4a2c38a5723",
"md5": "4d5ebf8c5fb10739a988c3ec07bcaad7",
"sha256": "81b7593bfdc805df84ae8cf7a00963bf79ef355cb0a74da3f41d38f0f19dab7c"
},
"downloads": -1,
"filename": "dsolve-0.0.16.tar.gz",
"has_sig": false,
"md5_digest": "4d5ebf8c5fb10739a988c3ec07bcaad7",
"packagetype": "sdist",
"python_version": "source",
"requires_python": null,
"size": 99077,
"upload_time": "2023-04-24T14:32:11",
"upload_time_iso_8601": "2023-04-24T14:32:11.670134Z",
"url": "https://files.pythonhosted.org/packages/fa/74/5b93be8027b5a2ca6b8d259d8e2b52bd671c3a9c05787cffb4a2c38a5723/dsolve-0.0.16.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2023-04-24 14:32:11",
"github": true,
"gitlab": false,
"bitbucket": false,
"github_user": "marcdelabarrera",
"github_project": "dsolve",
"travis_ci": false,
"coveralls": false,
"github_actions": false,
"lcname": "dsolve"
}