Add files for the whitebox tool

docs/index.rst

@@ -25,6 +25,7 @@ Furthermore, Privugger is equipped with a module to perform *automatic attacker
    tutorials/Tutorial.ipynb
    tutorials/Open-dp-Tutorial.ipynb
    tutorials/Governor.ipynb
+   tutorials/Duplicate.ipynb
 
 
 .. toctree::

docs/tutorials/output.py

@@ -0,0 +1,7 @@
+import numpy as np
+import pymc as pm
+import aesara.tensor as at
+
+
+def method(ages):
+    return ages.sum() / ages.size

docs/tutorials/randint.py

@@ -0,0 +1,12 @@
+import aesara
+import aesara.tensor as at
+import numpy as np
+
+
+@aesara.compile.ops.as_op(itypes=[at.lscalar], otypes=[at.lscalar])
+def method(arg0):
+
+    def randint(arg0):
+        y = np.random.randint(arg0)
+        return np.array(y)
+    return randint(arg0)

docs/tutorials/randint_program.py

@@ -0,0 +1,15 @@
+import numpy as np
+import pymc as pm
+import aesara.tensor as at
+import randint as ra
+
+
+def method(SECRET):
+    X = SECRET
+    pm.Deterministic('lign1', X)
+    var20 = pm.DiracDelta('var20', np.int64(100))
+    Y = ra.method(var20)
+    pm.Deterministic('lign2', Y)
+    x = X + Y
+    pm.Deterministic('lign3', x)
+    return x

docs/tutorials/test_multiple_deterministic.ipynb

@@ -0,0 +1,192 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import numpy as np\n",
+    "import pymc as pm\n",
+    "import os, sys\n",
+    "import random as rd\n",
+    "\n",
+    "sys.path.append(os.path.join(\"../../\"))\n",
+    "import privugger as pv"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def programme(SECRET):\n",
+    "    X = SECRET\n",
+    "    Y = np.random.randint(100)\n",
+    "    x = X + Y\n",
+    "    return x"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "SECRET = pv.Constant('SECRET', np.int64(42))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "ds = pv.Dataset(input_specs = [SECRET])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "program   = pv.Program('randint_program', dataset=ds, output_type=pv.Int, function=programme)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Multiprocess sampling (2 chains in 4 jobs)\n",
+      "CompoundStep\n",
+      ">Metropolis: [SECRET]\n",
+      ">Metropolis: [var20]\n"
+     ]
+    },
+    {
+     "data": {
+      "text/html": [
+       "\n",
+       "<style>\n",
+       "    /* Turns off some styling */\n",
+       "    progress {\n",
+       "        /* gets rid of default border in Firefox and Opera. */\n",
+       "        border: none;\n",
+       "        /* Needs to be in here for Safari polyfill so background images work as expected. */\n",
+       "        background-size: auto;\n",
+       "    }\n",
+       "    progress:not([value]), progress:not([value])::-webkit-progress-bar {\n",
+       "        background: repeating-linear-gradient(45deg, #7e7e7e, #7e7e7e 10px, #5c5c5c 10px, #5c5c5c 20px);\n",
+       "    }\n",
+       "    .progress-bar-interrupted, .progress-bar-interrupted::-webkit-progress-bar {\n",
+       "        background: #F44336;\n",
+       "    }\n",
+       "</style>\n"
+      ],
+      "text/plain": [
+       "<IPython.core.display.HTML object>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "text/html": [
+       "\n",
+       "    <div>\n",
+       "      <progress value='22000' class='' max='22000' style='width:300px; height:20px; vertical-align: middle;'></progress>\n",
+       "      100.00% [22000/22000 00:01&lt;00:00 Sampling 2 chains, 0 divergences]\n",
+       "    </div>\n",
+       "    "
+      ],
+      "text/plain": [
+       "<IPython.core.display.HTML object>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Sampling 2 chains for 1_000 tune and 10_000 draw iterations (2_000 + 20_000 draws total) took 5 seconds.\n"
+     ]
+    }
+   ],
+   "source": [
+    "trace = pv.infer(program, cores=4, draws=10000, method='pymc_whitebox', return_model=False, observation=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "<xarray.Dataset>\n",
+      "Dimensions:          (chain: 2, draw: 10000)\n",
+      "Coordinates:\n",
+      "  * chain            (chain) int64 0 1\n",
+      "  * draw             (draw) int64 0 1 2 3 4 5 ... 9994 9995 9996 9997 9998 9999\n",
+      "Data variables:\n",
+      "    SECRET           (chain, draw) int64 42 42 42 42 42 42 ... 42 42 42 42 42 42\n",
+      "    var20            (chain, draw) int64 100 100 100 100 100 ... 100 100 100 100\n",
+      "    lign1            (chain, draw) int64 42 42 42 42 42 42 ... 42 42 42 42 42 42\n",
+      "    lign2            (chain, draw) int64 58 73 94 7 83 90 ... 5 45 89 25 70 70\n",
+      "    lign3            (chain, draw) int64 100 115 136 49 125 ... 131 67 112 112\n",
+      "    randint_program  (chain, draw) int64 100 115 136 49 125 ... 131 67 112 112\n",
+      "Attributes:\n",
+      "    created_at:                 2022-08-26T13:47:31.467161\n",
+      "    arviz_version:              0.12.1\n",
+      "    inference_library:          pymc\n",
+      "    inference_library_version:  4.1.5\n",
+      "    sampling_time:              5.286848068237305\n",
+      "    tuning_steps:               1000\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(trace.posterior)"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3.10.5 ('privugger_environment')",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.6"
+  },
+  "orig_nbformat": 4,
+  "vscode": {
+   "interpreter": {
+    "hash": "ead01f34da02679ce3852bdbdb51f16b97d4b75626f761dacfb86d0793faee05"
+   }
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

privugger/data_structures/program.py

@@ -21,6 +21,13 @@ def __init__(self, name, dataset, output_type, function):
             self.program              = function
             self.observation          = None
             self.execute_observations = lambda a,b: None
+        elif dataset is None:
+            self.dataset              = None
+            self.output_type          = output_type
+            self.name                 = name
+            self.program              = function
+            self.observation          = None
+            self.execute_observations = lambda a,b: None
         else:
             raise ValueError("The dataset has to be of type privugger.Dataset")
 
@@ -84,28 +91,25 @@ def add_observation(self, constraints, precision=0.01):
 
         partial1 = lambda x: None
         partial2 = lambda x: None
-        if name in var_names or "output" in name:
-            if val1 != "" and cons1 != "":
-                v1 = float(val1) if "." in val1 else int(val1)
-                partial1 = self._unwrap_constrain(v1, cons1, precision)
-                print(partial1)
-
-            if val2 != "" and cons2 != "":
-                v2 = float(val2) if "." in val2 else int(val2)
-                partial2 = self._unwrap_constrain(v2, cons2, precision, i=1)
-
-            def inner(prior, output):
-                if name in var_names:
-                    idx = var_names.index(name)
-                    distribution = prior[idx]
-                else:
-                    distribution = output
-                partial1(distribution)
-                partial2(distribution)
-            self.execute_observations = inner
-            return None # to avoid having a return value
-        else:
-            raise ValueError("Observation was not known. Make sure that the name is part of the names in privugger.Datastructure")
+        if val1 != "" and cons1 != "":
+            v1 = float(val1) if "." in val1 else int(val1)
+            partial1 = self._unwrap_constrain(v1, cons1, precision)
+            print(partial1)
+
+        if val2 != "" and cons2 != "":
+            v2 = float(val2) if "." in val2 else int(val2)
+            partial2 = self._unwrap_constrain(v2, cons2, precision, i=1)
+
+        def inner(prior, output):
+            if name in var_names:
+                idx = var_names.index(name)
+                distribution = prior[idx]
+            else:
+                distribution = output
+            partial1(distribution)
+            partial2(distribution)
+        self.execute_observations = inner
+        return None # to avoid having a return value
 
 
     def _unwrap_constrain(self, value, cons, precision, i=0):

privugger/transformer/PyMC3/type_decoration.py

@@ -107,6 +107,9 @@ def lift(self, program, decorators):
                 file.write(res)
             tree = ast.parse(open("temp.py").read())
 
+            #first_deterministic = ast.Expr(value=ast.Call(func=ast.Attribute(value=ast.Name(id='pm', ctx=ast.Load()), attr='Deterministic', ctx=ast.Load()), args=[ast.Constant(value='rd_countries'), ast.Name(id='rd_countries', ctx=ast.Load())], keywords=[]))
+            #tree.body[0].body.insert(6, first_deterministic)
+
             os.remove("temp.py")
 
         #if("lambda")
@@ -144,7 +147,7 @@ def translate_type(self, p_type):
             return 'dscalar'
 
         elif(p_type == 'int'):
-            return 'wscalar'
+            return 'lscalar'
 
         elif(p_type == 'VectorI'):
             return 'lvector'
@@ -415,7 +418,7 @@ def wrap_with_theano_import(self, program):
         aesara_import = ast.Import(names=[ast.alias(name='aesara', asname=None)])
         aesara_tensor_import = ast.Import(names=[ast.alias(name='aesara.tensor', asname='at')])
         numpy_import = ast.Import(names=[ast.alias(name='numpy', asname='np')])
-        new_program = ast.Module(body=[aesara_import, aesara_tensor_import,numpy_import,  program])
+        new_program = ast.Module(body=[aesara_import, aesara_tensor_import, numpy_import, program])
 
         return new_program