smbinterp/interp/grid/__init__.py

import sys
import re
from   collections import defaultdict
import inspect

import numpy as np
import scipy.spatial

from interp.baker import run_baker
from interp.tools import exact_func, log
from simplex      import face, contains
from smcqdelaunay import *


class grid(object):
  def __init__(self, verts = None, q = None):
    """
      this thing eats two pre-constructed arrays of floats:

      verts = array of arrays (if passed in, will convert to numpy.array)
              [
                [x0,y0],
                [x1,y1], ...
              ]

      q     = array (1D) of physical values
    """

    if verts:
      self.verts = np.array(verts)
      self.tree           = scipy.spatial.KDTree(self.verts)
    if q:
      self.q = np.array(q)

    self.faces          = {}
    self.faces_for_vert = defaultdict(list)

  def get_containing_simplex(self, X):
    if not self.faces:
      raise Exception("face connectivity is not set up")

    # get closest point
    (dist, indicies) = self.tree.query(X, 2)
    closest_point = indicies[0]

    log.debug('X: %s' % X)
    log.debug('point index: %d' % closest_point)
    log.debug('actual point %s' % self.verts[closest_point])
    log.debug('distance = %0.4f' % dist[0])

    simplex = None
    checked_faces = []
    faces_to_check = self.faces_for_vert[closest_point]

    attempts = 0
    while not simplex and faces_to_check:
      attempts += 1

      # if attempts > 20:
      #   raise Exception("probably recursing to many times")

      cur_face = faces_to_check.pop(0)
      checked_faces.append(cur_face)

      if cur_face.contains(X, self):
        simplex = cur_face
        continue

      new_facest = []
      for neighbor in cur_face.neighbors:
        if (neighbor not in checked_faces) and (neighbor not in faces_to_check):
          faces_to_check.append(neighbor)

    if not simplex:
      raise AssertionError('no containing simplex found')

    R = self.create_mesh(simplex.verts)

    log.debug('total attempts before finding simplex: %d' % attempts)
    return R

  def create_mesh(self, indicies):
    """
      this function takes a list of indicies, and then creates
      and returns a grid object (collection of verts and q).

      note: the input is indicies, the grid contains verts
    """
    p = [self.verts[i]  for i in indicies]
    q = [self.q[i]      for i in indicies]
    return grid(p, q)

  def get_simplex_and_nearest_points(self, X, extra_points = 3, simplex_size = 3):
    """
      this returns two grid objects: R and S.

      R is a grid object that is supposedly a containing simplex
        around point X (TODO: it tends not to be)

      S is S_j from baker's paper : some verts from all point that are not the simplex
    """
    log.debug("extra verts: %d" % extra_points)
    log.debug("simplex size: %d" % simplex_size)

    r_mesh = self.get_containing_simplex(X)

    # and some UNIQUE extra verts
    (dist, indicies) = self.tree.query(X, simplex_size + extra_points)
    log.info("extra indicies: %s" % indicies)

    unique_indicies = []
    for index in indicies:
      close_point_in_R = False
      for rvert in r_mesh.verts:
        if all(rvert == self.verts[index]):
          close_point_in_R = True
          break

      if not close_point_in_R:
        unique_indicies.append(index)
      else:
        log.debug('throwing out %s: %s' % (index, self.verts[index]))

    log.debug("indicies: %s" % indicies)
    log.debug("unique indicies: %s" % unique_indicies)
    s_mesh = self.create_mesh(unique_indicies)

    return (r_mesh, s_mesh)

  def run_baker(self, X, extra_points = 3, order = 2):
    (R, S) = self.get_simplex_and_nearest_points(X, extra_points)
    answer = run_baker(X, R, S, order)
    return answer

  def __str__(self):
    r = ''
    assert( len(self.verts) == len(self.q) )
    for c, i in enumerate(zip(self.verts, self.q)):
      r += "%d vert(%s): q(%0.4f)" % (c,i[0], i[1])
      face_str = ", ".join([str(f.name) for f in self.faces_for_vert[c]])
      r += "  faces: [%s]" % face_str
      r += "\n"
    if self.faces:
      for v in self.faces.itervalues():
        r += "%s\n" % v
    return r



class delaunay_grid(grid):
  face_re = re.compile(r'''
                            -\s+(?P<face>f\d+).*?
                            vertices:\s(?P<verts>.*?)\n.*?
                            neighboring\s faces:\s+(?P<neigh>[\sf\d]*)
                         ''', re.S|re.X)

  point_re = re.compile(r'''
                            -\s+(?P<point>p\d+).*?
                            neighbors:\s+(?P<neigh>[\sf\d]*)
                         ''', re.S|re.X)

  vert_re  = re.compile(r'''
                            (p\d+)
                         ''', re.S|re.X)

  def __init__(self, verts, q):
    grid.__init__(self, verts,q)


  def get_containing_simplex(self, X):
    if not self.faces:
      self.construct_connectivity()

    # get closest point
    (dist, indicies) = self.tree.query(X, 2)
    closest_point = indicies[0]

    log.debug('X: %s' % X)
    log.debug('point index: %d' % closest_point)
    log.debug('actual point %s' % self.verts[closest_point])
    log.debug('distance = %0.4f' % dist[0])

    simplex = None
    checked_faces = []
    faces_to_check = self.faces_for_vert[closest_point]

    attempts = 0
    while not simplex and faces_to_check:
      attempts += 1
      # if attempts > 20:
      #   raise Exception("probably recursing to many times")
      cur_face = faces_to_check.pop(0)
      checked_faces.append(cur_face)

      if cur_face.contains(X, self):
        simplex = cur_face
        continue

      new_facest = []
      for neighbor in cur_face.neighbors:
        if (neighbor not in checked_faces) and (neighbor not in faces_to_check):
          faces_to_check.append(neighbor)

    if not simplex:
      raise AssertionError('no containing simplex found')

    R = self.create_mesh(simplex.verts)

    log.debug('total attempts before finding simplex: %d' % attempts)
    return R


  def get_simplex_and_nearest_points(self, X, extra_points = 3, simplex_size = 3):
    """
      this returns two grid objects: R and S.

      R is a grid object that is supposedly a containing simplex
        around point X (it tends not to be)

      S is S_j from baker's paper : some verts from all point that are not the simplex
    """
    log.debug(inspect.stack()[1][3])
    log.debug("extra verts: %d" % extra_points)
    log.debug("simplex size: %d" % simplex_size)

    r_mesh = self.get_containing_simplex(X)
    # log.debug("R:\n%s" % r_mesh)

    # and some UNIQUE extra verts
    (dist, indicies) = self.tree.query(X, simplex_size + extra_points)

    unique_indicies = []
    for index in indicies:
      if self.verts[index] not in r_mesh.verts:
        unique_indicies.append(index)

    log.debug("indicies: %s" % ",".join([str(i) for i in indicies]))
    log.debug("indicies: %s" % ",".join([str(i) for i in unique_indicies]))
    s_mesh = self.create_mesh(unique_indicies)# indicies[simplex_size:])

    # TODO: eventually remove this test:
    for point in s_mesh.verts:
      if point in r_mesh.verts:
        log.error("ERROR")
        log.error("\n%s\nin\n%s" % (point, r_mesh))
        raise Exception("repeating point S and R")

    return (r_mesh, s_mesh)

  def get_points_conn(self, X):
    """
      this returns two grid objects: R and S.

      this function differes from the get_simplex_and_nearest_points
      function in that it builds up the extra verts based on
      connectivity information, not just nearest-neighbor.
      in theory, this will work much better for situations like
      verts near a short edge in a boundary layer cell where the
      nearest verts would all be colinear

      also, it guarantees that we find a containing simplex

      R is a grid object that is the (a) containing simplex around point X
      S is a connectivity-based nearest-neighbor lookup, limited to 3 extra verts
    """
    if not self.faces:
      self.construct_connectivity()

    # get closest point
    (dist, indicies) = self.tree.query(X, 2)

    simplex = None
    for face in self.faces_for_vert[indicies[0]]:
      if face.contains(X, self):
        simplex = face
        break

    if not simplex:
      raise AssertionError('no containing simplex found')

    # self.create_mesh(simplex.verts)
    R = self.get_containing_simplex(X)

    s = []
    for c,i in enumerate(simplex.neighbors):
      s.extend([guy for guy in i.verts if not guy in simplex.verts])
    S = self.create_mesh(s)

    return R, S

  def run_baker(self, X, extra_points = 3, order = 2):
    answer = None

    try:
      (R, S) = self.get_simplex_and_nearest_points(X)
      if not contains(X, R.verts):
        raise Exception("run_baker with get_simplex_and_nearest_points returned non-containing simplex")
      answer = run_baker(X, R, S, order)
    except Exception, e:
      log.error("caught error: %s, trying with connectivity-based mesh" % e)
      (R, S) = self.get_points_conn(X)
      answer = run_baker(X, R, S, order)

    return answer



  def construct_connectivity(self):
    """
      a call to this method prepares the internal connectivity structure.

      this is part of the __init__ for a rect_grid, but can be called from any grid object
    """
    log.debug('start')
    qdelaunay_string = get_qdelaunay_dump_str(self)
    face_to_faces = []
    for matcher in grid.face_re.finditer(qdelaunay_string):
      d = matcher.groupdict()

      face_name          = d['face']
      verticies           = d['verts']
      neighboring_faces  = d['neigh']

      cur_face = face(face_name)
      self.faces[face_name] = cur_face

      for v in grid.vert_re.findall(verticies):
        vertex_index = int(v[1:])
        cur_face.add_vert(vertex_index)
        self.faces_for_vert[vertex_index].append(cur_face)

      nghbrs = [(face_name, i) for i in  neighboring_faces.split()]
      face_to_faces.extend(nghbrs)

    for rel in face_to_faces:
      if rel[1] in self.faces:
        self.faces[rel[0]].add_neighbor(self.faces[rel[1]])

    log.debug('end')