Merge pull request #88 from deepskies/dev

functionality to specify which cls to output + other (param/option name) updates for clarity

deepcmbsim/camb_power_spectrum.py

@@ -63,7 +63,7 @@ def get_noise(self):
             provides noise for the TT power spectrum and the polarization power spectra;
             shape is (2, max_l_use)
         """
-        if self.UserParams['noise_type'] == 'white':
+        if self.UserParams['noise_type'] == 'detector-white':
             t_noise = noise.detector_white_noise(self.UserParams['noise_uKarcmin'], self.UserParams['beamfwhm_arcmin'], self.max_l_use,
                                                  TT=True)
             eb_noise = noise.detector_white_noise(self.UserParams['noise_uKarcmin'], self.UserParams['beamfwhm_arcmin'], self.max_l_use,
@@ -72,7 +72,7 @@ def get_noise(self):
         elif self.UserParams['noise_type'] is None:
             return np.zeros((2, self.max_l_use))
         else:
-            print("only detector white noise is currently implemented, via `noise_type = 'white'` in `user_config.yaml`")
+            print("only detector white noise is currently implemented, via `noise_type = 'detector-white'` in `user_config.yaml`")
             return np.zeros((2, self.max_l_use))
 
     def get_cls(self, save_to_dict=None, user_params=True):
@@ -98,29 +98,51 @@ def get_cls(self, save_to_dict=None, user_params=True):
         # main calculation: https://camb.readthedocs.io/en/latest/camb.html#camb.get_results
         results = camb.get_results(self.CAMBparams)
 
+        outdict = { 'l': range(self.max_l_use + 1) }
+        if self.UserParams['cls_to_output'] == 'all':
+            cls_needed = ['clTT', 'clEE', 'clBB', 'clTE', 'clPP', 'clPT', 'clPE']
+        else:
+            cls_needed = self.UserParams['cls_to_output']
+
         # https://camb.readthedocs.io/en/latest/results.html#camb.results.CAMBdata.get_total_cls
-        tt, ee, bb, te = results.get_total_cls(raw_cl=self.normalize_cls, CMB_unit=self.TT_units)[:self.max_l_use + 1].T
-        if self.UserParams['noise_type'] is not None:
-            _noise = self.get_noise()
-            tt += _noise[0]
-            ee += _noise[1]
-            bb += _noise[1]
-            te += _noise[1]
+        if ('clTT' in cls_needed) or ('clEE' in cls_needed) or ('clEB' in cls_needed) or ('clTE' in cls_needed):
+            # need to run things to get one/some/all of tt, ee, bb, te
+            tt, ee, bb, te = results.get_total_cls(raw_cl=self.normalize_cls, CMB_unit=self.TT_units)[:self.max_l_use + 1].T
+            # add noise
+            if self.UserParams['noise_type'] is not None:
+                _noise = self.get_noise()
+                tt += _noise[0]
+                ee += _noise[1]
+                bb += _noise[1]
+                te += _noise[1]
+            # now add to outdict
+            for key in ['clTT', 'clEE', 'clBB', 'clTE']:
+                if key in cls_needed:
+                    if key == 'clTT':
+                        outdict[key] = tt
+                    elif key == 'clEE':
+                        outdict[key] = ee
+                    elif key == 'clBB':
+                        outdict[key] = bb
+                    elif key == 'clTE':
+                        outdict[key] = te
+                    else:
+                        raise ValueError('somethings wrong.')
 
         # https://camb.readthedocs.io/en/latest/results.html#camb.results.CAMBdata.get_lens_potential_cls
-        pp, pt, pe = results.get_lens_potential_cls(raw_cl=self.normalize_cls)[:self.max_l_use + 1].T
-        lvals = range(self.max_l_use + 1)
-
-        outdict = {
-            'l': lvals,
-            'clTT': tt,
-            'clEE': ee,
-            'clBB': bb,
-            'clTE': te,
-            'clPP': pp,
-            'clPT': pt,
-            'clPE': pe
-        }
+        if ('clPP' in cls_needed) or ('clPT' in cls_needed) or ('clPE' in cls_needed):
+            pp, pt, pe = results.get_lens_potential_cls(raw_cl=self.normalize_cls)[:self.max_l_use + 1].T
+            # now add to outdict
+            for key in ['clPP', 'clPT', 'clPE']:
+                if key in cls_needed:
+                    if key == 'clPP':
+                        outdict[key] = pp
+                    elif key == 'clPT':
+                        outdict[key] = pt
+                    elif key == 'clPE':
+                        outdict[key] = pe
+                    else:
+                        raise ValueError('somethings wrong.')
 
         if bool(self.UserParams["verbose"]):
             time_end = dt.now()

deepcmbsim/cl_plotting.py

@@ -15,8 +15,8 @@ class flatmap:
         number of pixels on each side of a flat map
     degrees : float
         number of degrees on each side of a flat map
-    seed : int, default -1
-        random seed to provide `namaster` (potential bug in `namaster`, does not always have the
+    namaster_seed : int, default -1
+        random namaster_seed to provide `namaster` (potential bug in `namaster`, does not always have the
         behavior that is described at https://namaster.readthedocs.io/en/latest/pymaster.html
     cl_dict : dict, optional
         dictionary of power spectra. Necessary to use the primary method of this class.
@@ -40,7 +40,7 @@ class flatmap:
         function for plotting a realization of a flat-sky map with the given number of pixels
         representing the given size of sky patch based on the provided power spectra
     """
-    def __init__(self, pixels, degrees, seed = -1, cl_dict = None):
+    def __init__(self, pixels, degrees, namaster_seed = -1, cl_dict = None):
         """
         initialize a map-making class
         Parameters
@@ -49,32 +49,32 @@ def __init__(self, pixels, degrees, seed = -1, cl_dict = None):
             number of pixels on each side of a flat map
         degrees : float
             number of degrees on each side of a flat map
-        seed : int, default -1
-            random seed to provide `namaster` (potential bug in `namaster`, does not always have the
+        namaster_seed : int, default -1
+            random namaster_seed to provide `namaster` (potential bug in `namaster`, does not always have the
             behavior that is described at https://namaster.readthedocs.io/en/latest/pymaster.html
         cl_dict : dict, optional
             dictionary of power spectra. Necessary to use the primary method of this class.
             If not specified, can still use the hidden mapping function.
         """
         self.pixels, self.degrees = pixels, degrees  # pixels on each side, degrees on each side
-        self.seed, self.cl_dict = seed, cl_dict # random seed if applicable, dictionary of CLs if you want to include them
+        self.namaster_seed, self.cl_dict = namaster_seed, cl_dict # random namaster_seed if applicable, dictionary of CLs if you want to include them
         self.reso = self.degrees / self.pixels
         self.nx, self.lx, self.lx_rad, self.ny, self.ly, self.ly_rad = [int(self.pixels), int(self.degrees), self.degrees*np.pi/180]*2
         self.ticks, self.labels = np.linspace(0, self.pixels, self.degrees+1), np.arange(0, self.degrees+1, dtype=float)
 
-    def _flatmap(self, cl_array, spin_array = [0], seed = None):
-        seed = seed if seed is not None else self.seed
-        return nmt.synfast_flat(self.nx, self.ny, self.lx_rad, self.ly_rad, np.array([x for x in cl_array]), spin_array, seed = seed)
+    def _flatmap(self, cl_array, spin_array = [0], namaster_seed = None):
+        namaster_seed = namaster_seed if namaster_seed is not None else self.namaster_seed
+        return nmt.synfast_flat(self.nx, self.ny, self.lx_rad, self.ly_rad, np.array([x for x in cl_array]), spin_array, namaster_seed = namaster_seed)
 
-    def flatmap(self, maps_out, seed = None):
+    def flatmap(self, maps_out, namaster_seed = None):
         """
 
         Parameters
         ----------
         maps_out : ["T", "E", "B", "P", "TT", "EE", "BB", "TE", "PP", "PT", "PE", "clTT", "clEE", "clBB", "clTE", "clPP", "clPT", "clPE", "TEB", "TQU"]
             map(s) that you would like to produce
-        seed : int, optional
-            random seed to provide `namaster` (potential bug in `namaster`, does not always have the
+        namaster_seed : int, optional
+            random namaster_seed to provide `namaster` (potential bug in `namaster`, does not always have the
             behavior that is described at https://namaster.readthedocs.io/en/latest/pymaster.html
         Returns
         -------
@@ -98,6 +98,6 @@ def flatmap(self, maps_out, seed = None):
             else:
                 print("not a valid map specification")
                 return None
-            return self._flatmap(cl_arr, spin_arr, seed=seed)
+            return self._flatmap(cl_arr, spin_arr, namaster_seed=namaster_seed)
         else:
             print("if you don't want to restrict to a `cl_dict` dictionary, use `self._flatmap` instead")

deepcmbsim/settings/user_config.yaml

@@ -3,14 +3,18 @@ FORCAMB:  # these simply overwrite any values in BASECAMBPARAMS
 ITERABLES:  # these provide iterables that can overwrite their corresponding values in BASECAMBPARAMS in loop settings
   InitPower.r : [0.01, 0.1, 3]  # for nested CAMBparams attributes, use the dot structure
   Alens : [.8, 1.2, 3]
-seed: 0
+namaster_seed: 0   # seed for the map realization using namaster
 verbose: 1
 normalize_cls: False #raw_cl – return Cl rather than l*(l+1)*Cl/2π (Cl alone is not conventional)
 TT_units: muK #return uK**2 units for the TT, EE, BB, TE Cls
 outfile_dir: "outfiles"
-noise_type: "white"
+noise_type: "detector-white"    # only option for now; adds detector white noise to the output; see noise.detector_white_noise for more details.
 noise_uKarcmin: 5  # noise level in uK*arcmin
 beamfwhm_arcmin: 3  # size of beam in arcmin
 extra_l: 300
 max_l_use: 10000  # max_l_use will differ from max_l and max_l_tensor by "extra_l" because
-# according to the CAMB documentation errors affect the last "100 or so" multipoles
+# according to the CAMB documentation errors affect the last "100 or so" multipoles
+# ---
+# decide on what to output; either 'all' or a subset list of
+# ['clTT', 'clEE', 'clBB', 'clTE', 'clPP', 'clPT', 'clPE']
+cls_to_output: 'all'