# ---------------------------------------------------------------------------
# morans_I_plus_general_G_variable_distance_better_flow_control.py
# Created on: Sat Feb 25 2006 09:29:17 AM
#   (generated by ArcGIS/ModelBuilder)
# ---------------------------------------------------------------------------
## This version is actually running variable fixed distance in increments of 1 km.

# Import system modules
import sys, string, os, win32com.client

# Create the Geoprocessor object
gp = win32com.client.Dispatch("esriGeoprocessing.GpDispatch.1")

# Load required toolboxes...
gp.AddToolbox("C:/Program Files/ArcGIS/ArcToolbox/Toolboxes/Spatial Statistics Tools.tbx")

weed = ["mAGRRE","mAGSSTAGST","mAMAPOAMAR","mANTCO","mAVESAAVEF","mBROMUSspp","mCAPBP","mCatsearli","mCIRAR","mCONAR","mDACGL_log"] # 0 to 10
weed += ["mDAUCA","mECNOR","mEQUAR","mFESAR_log","mGALAP","mHOLLA","mHOLMO","mKICEL","mLACSE","mLOLMU_log","mLOLPE_log","mLOWLNDCUD"] # 11 to 22
weed += ["mMATMT","mPOAAN_log","mPOAPR","mPOATR","mPOLPE","mRUBDI","mRUMOBRUMC","mSENVU","mSINAR","mSONAR","mTYPAR","mVLPMY","mWheat_log"] # 23 to 35
weed += ["wAGSSTAGST","wDACGL","wFESAR","wLOLMU","wLOLPE","wPOAPR","wPOATR"] #36 to 42

# started out with only a few estimates of likely values for the peak autocorrelation and maximum range.
# This batch file was only run using the original 0,1,2,3,4 weed severity conversion scale, not the newer one.

init_dist_peak_MI_prob = [9.0,1.0,0.0,0.0,2.5,2.0,0.0,0.0,2.5,4.0,0.0,0.0,0.0,0.0,0.0,0.0,21.0,8.0,0.0,0.0,8.0,0.0,0.0,0.0,9.0,0.0,4.0,0.0,0.0,0.0] # 0 to 29
init_dist_peak_MI_prob += [0.0,0.0,0.0,0.0,0.0,0.0,1.0,0.0,0.0,9.0,0.0,0.0,4.0] # 30 to 42 - values are in miles

init_dist_lose_MI_prob = [31.0,25.0,0.0,0.0,25.0,26.0,0.0,0.0,28.0,20.0,0.0,0.0,0.0,0.0,0.0,0.0,50.0,47.0,0.0,0.0,26.0,0.0,0.0,0.0,28.0,0.0,20.0,0.0,0.0,0.0] # 0 to 29
init_dist_lose_MI_prob += [0.0,0.0,0.0,0.0,0.0,0.0,25.0,0.0,0.0,23.0,0.0,0.0,20.0] # 30 to 42 - values are in miles

LinnBoundaries = "C:\\DataFiles\\OSUSeedCert\\SpatialStats\\LinnBoundaries.shp"
gp_count = 0
summary_file_name = "C:\\DataFiles\\OSUSeedCert\\SpatialStats\\MI_distance_summary.txt"
summary_file = open(summary_file_name, "w") # write the observed peak range and outer bound of significance range to a file - will rewrite existing file if opened in "w" mode
GGsummary_file_name = "C:\\DataFiles\\OSUSeedCert\\SpatialStats\\GG_distance_summary.txt"
GGsummary_file = open(GGsummary_file_name, "w") # write the observed peak range and outer bound of significance range to a file - will rewrite existing file if opened in "w" mode

# algorithm to find distance of maximum MI Z-score
first_try = 0
using_defaults = 0
number_done = 0 # need to increment this value in future restarts of program    
for weed_num in range(number_done,43):  # setting this to 1 for now, but does eventually need to be 43
    test_distance = init_dist_peak_MI_prob[weed_num]
    lose_sig_test_dist = init_dist_lose_MI_prob[weed_num]
    if test_distance == 0.0:
        test_distance = 6.67 # average of 12 already nearly done
        first_try = 1
        using_defaults = 1
    test_distance *= 1.609344 # convert into kilometers
    test_distance = int(test_distance) # convert to integers in kilometers for use in dictionary
    if lose_sig_test_dist == 0.0:
        lose_sig_test_dist = 29.08 # of 12 already partially done
    lose_sig_test_dist *= 1.609344 # convert into kilometers
    lose_sig_test_dist = int(lose_sig_test_dist) # convert to integers in kilometers for use in dictionary    
    weed_name = weed[weed_num]
    MI_results_name = "C:\\DataFiles\\OSUSeedCert\\SpatialStats\\MI_"+weed_name+"_variable_distance_results_for_max_Z_score.txt"
    MI_results_file = open(MI_results_name,"a+")
    print MI_results_name+" "+weed_name
    print "lose sig test dist from program data list = "+str(lose_sig_test_dist)
    peak_reached = 0
    floor_reached = 0
    Z_scores_observed = {-1:0.0,0:0.0} # dummy values of Z = 0,0 for distance of influence = -1,0 ## -1 needed to give full range around possible solution of 1000 m
    # now load any existing data from previously written file
    already_made = MI_results_file.readline()
    max_Z = 0
    temp_peak = 0
    while (len(already_made) > 0) and (already_made != "\n"): # skip this if nothing is in file
        #convert line into dictionary entry
        space1 = already_made.find(" ")
        space2 = already_made.find(" ",space1 + 1)
        space3 = already_made.find(" ",space2 + 1)
        space4 = already_made.find(" ",space3 + 1)
        final_separator = already_made.find("; ",space4 + 1)
        existing_distance = int(already_made[space1 + 1:space2])
        existing_zscore = float(already_made[final_separator + 2:])
        if existing_zscore < 0:
            existing_zscore = 0
        Z_scores_observed[existing_distance] = existing_zscore
        # print str(existing_distance)+" "+str(existing_zscore)
        already_made = MI_results_file.readline() # keep reading in more lines until while loop stops
        first_try = 0 # have some results in stored text file
        if existing_zscore > max_Z:
            max_Z = existing_zscore
            temp_peak = existing_distance
    print Z_scores_observed
    print str(max_Z)+" = max Z-score in text file "+str(temp_peak)+" = distance in km"
    MI_results_file.close()
    MI_results_file = open(MI_results_name,"a")
    if temp_peak > 0:
        if test_distance == 10: # change to better value if we've already run program for this weed and written data to text file
            test_distance = temp_peak
        if test_distance != 10:
            print "not changing test distance from "+str(test_distance)+" km in local data to "+str(temp_peak)+" in results stored in text file"
            # leave test_distance alone unless we add more code 
    while peak_reached == 0:
        for this_shift in range(-2,3): # plus and minus 1 and 2 km from test_distance except during first_try on new weed
            if first_try == 0:
                shifting = this_shift
            else:
                shifting = (2 + this_shift) * 10 # generate wide range 10,20,30,40,50 km if no preliminary search for peak
            local_distance = test_distance + shifting
            print "local distance value is "+str(local_distance)
            if local_distance <= 0:
                local_distance = 1 # do not run at less than 1000 m range
            if not Z_scores_observed.has_key(local_distance): # only run if this distance hasn't already been done
                # Local variables...        
                Index = "0"
                ZScore = "0"
                MI_results_file.write(weed_name+" "+str(local_distance)+" in kilometers ")  
                param_distance = 3280.839895 * local_distance  # conversion factor is feet in a kilometer    
                # Process: Spatial Autocorrelation (Morans I)...
                curr_results = gp.SpatialAutocorrelation_stats(LinnBoundaries, weed_name, "false", "Fixed Distance Band", "Euclidean Distance", "None", param_distance, "")
                gp_count +=1
                MI_results_file.writelines(curr_results)
                MI_results_file.write("\n")
                MI_results_file.flush()
                myZ = float(curr_results[curr_results.find(";")+1:])
                # print " Z-score = "+str(myZ)+" for distance in km of "+str(local_distance)
                if myZ < 0:
                    myZ = 0 # do need keep negative Z-scores as they make no sense in our context and might break the search procedure for maximum Z-score
                Z_scores_observed[local_distance] = myZ
        values_so_far = Z_scores_observed.values()
        max_value = max(values_so_far)
        distances_so_far = Z_scores_observed.keys()
        peak = 0
        for where_is_peak in distances_so_far:
            if Z_scores_observed[where_is_peak] == max_value:
                peak = where_is_peak # this logic will get last value of several duplicate maximum values, but duplicates are very unlikely!!!  
        contiguity = 1
        for testing_contiguity in range(-2,3): # testing whether data exists in a contiguous range of -2,-1,+1,+2 around observed peak
            this_distance = peak + testing_contiguity
            if not Z_scores_observed.has_key(this_distance):
                contiguity = 0       
        if contiguity == 1:
            peak_reached = 1
        else: # still haven't found the peak, need to keep moving up or down in distance
            test_distance = peak
            first_try = 0
            print "new test distance in kilometers = "+str(test_distance)
    print "peak in MI Z-scores found at "+str(peak)+" with Z-value = "+str(Z_scores_observed[peak])
    summary_file.write(weed_name+"                 peak MI distance = "+str(peak)+" kilometers with Z-score = "+str(Z_scores_observed[peak])+"\n")
    summary_file.write(str(gp_count)+" gp calls this run."+"\n")
    summary_file.flush()

    # now trying to find distance for minimum Z-score > 1.96 with adjacent larger distances non-significant and shorter one more significant
    print "now running the find distance at which significance is first lost algorithm"
    new_min = Z_scores_observed[peak] # start new_min as high as any Z-score calculated so far
    if new_min < 1.96:
        print "warning, will robinson, danger!!!! - maximum observed Z-score was less than that required for 5% significance level"
    
    floor_reached = 0 # forgot to give this a default value    
    shortest_found = 0
    shortest_ns = max(distances_so_far) # a safe default value!
    if shortest_ns < lose_sig_test_dist:
        shortest_ns = lose_sig_test_dist # fixes case of peak search not loading dictionary with distance as large as initial estimate of lose_sig_test_dist
    if (lose_sig_test_dist == 46) and (using_defaults == 1): # the default value loaded for the no-data case - resetting to better behaved starting values
        lose_sig_test_dist -= 2 # drop distance slightly from tentative loss of significance distance
    distances_so_far = Z_scores_observed.keys()
    distances_so_far.sort()
    for where_is_first_floor in distances_so_far:
        if (Z_scores_observed[where_is_first_floor] < 1.96) and (where_is_first_floor > 0) and (shortest_found == 0)\
           and (where_is_first_floor > peak):
            shortest_ns = where_is_first_floor
            shortest_found = 1 # this code finds the first distance dropping below significance if one already exists in data
            lose_sig_test_dist = shortest_ns - 2 # shift back to try to catch passage from signif. to NS
            #if Z_scores_observed[where_is_first_floor] <= 0:
                #lose_sig_test_dist = shortest_ns - 2 # shift back even further if Z-score was below zero!
    if (shortest_found == 0) and (lose_sig_test_dist == 44) and (using_defaults == 1) and (lose_sig_test_dist != (peak + 5)): 
        lose_sig_test_dist = peak + 5 # NS values were not present and using default average for initial distance target 

    floor = shortest_ns # new control variable
    old_floor = floor
    stuck = 0
    escape_count = 0
    while floor_reached == 0:
        escape_count += 1
        if escape_count > 200:
            print "trouble at line 141"
            break
        for this_shift in range(-2,3): # plus and minus 1 and 2 km from lose_sig_test_dist
            escape_count += 1
            if escape_count > 200:
                print "trouble at line 143"
                break
            local_distance = lose_sig_test_dist + this_shift
            print "local distance value is "+str(local_distance)
            if local_distance <= 0:
                local_distance = 1 # do not run at less than 1000 m range
            if not Z_scores_observed.has_key(local_distance): # only run if this distance hasn't already been done
                print "in the floor reach -2,-1,0,1,2 loop"
                # Local variables...        
                Index = "0"
                ZScore = "0"
                MI_results_file.write(weed_name+" "+str(local_distance)+" in kilometers ")  
                param_distance = 3280.839895 * local_distance  # conversion factor is feet in a kilometer    
                # Process: Spatial Autocorrelation (Morans I)...
                curr_results = gp.SpatialAutocorrelation_stats(LinnBoundaries, weed_name, "false", "Fixed Distance Band", "Euclidean Distance", "None", param_distance, "")
                gp_count += 1
                MI_results_file.writelines(curr_results)
                MI_results_file.write("\n")
                MI_results_file.flush()
                myZ = float(curr_results[curr_results.find(";")+1:])
                # print " Z-score = "+str(myZ)+" for distance in km of "+str(local_distance)
                if myZ < 0:
                    myZ = 0 # do need keep negative Z-scores as they make no sense in our context and might break the search procedure for maximum Z-score
                Z_scores_observed[local_distance] = myZ
        values_so_far = Z_scores_observed.values()
        print values_so_far # dubugging printout
        old_min = new_min
        #for temp_Z in values_so_far: # need to track non-zero minimums
            # need to limit next test to just cases where distance > peak
        #    if (temp_Z < new_min) and (temp_Z >= 1.96): # using loss of 5% significance level for floor in finding maximum range of spatial autocorrelation
        #        new_min = temp_Z

############################# status condition ########################

        distances_so_far = Z_scores_observed.keys()
        testing_ns = max(distances_so_far) 
        consecutive_peak_to_curr_count = 0    
        peak_to_curr_gt196_count = 0   
        contiguity_peak_to_curr = 1 
        two_in_a_row_Z_lt_196 = 0
        pos_first_ns = testing_ns + 1 # safe default value
        pos_two_in_a_row = 0 # safe default
        least_dist_past_peak_Z_lt_196 = 0 # safe default
        distances_so_far.sort() 

        for test_contiguity in range(peak,lose_sig_test_dist+1):
            if not(Z_scores_observed.has_key(test_contiguity)):
                contiguity_peak_to_curr = 0 # contiguity does not yet exist in range peak to lose_sig_test_dist

        for where_is in distances_so_far: # loop thru distances in order, focusing on conditions past the peak
            if Z_scores_observed.has_key(where_is + 1) and Z_scores_observed.has_key(where_is + 2): # verify data exists at next two step on out
                if (Z_scores_observed[where_is] > 1.96) and (Z_scores_observed[where_is + 1] < 1.96) and (Z_scores_observed[where_is + 2] < 1.96) and (floor_reached == 0) and (where_is >= peak):
                    floor_reached = 1
                    floor = where_is # this completes things that needed to be done once floor location has been found

            if (where_is > peak) and (where_is <= lose_sig_test_dist):
                consecutive_peak_to_curr_count += 1
                if (Z_scores_observed[where_is] > 1.96):
                    peak_to_curr_gt196_count += 1
                if (Z_scores_observed[where_is] < 1.96) and (two_in_a_row_Z_lt_196 == 1) and (where_is == pos_first_ns + 1):
                    two_in_a_row_Z_lt_196 = 2
                    pos_two_in_a_row = where_is
                    # this is trigger for Case 5
                else: # reset count of two in a row NS Z-scores to 0
                    if (two_in_a_row_Z_lt_196 == 1):
                        two_in_a_row_Z_lt_196 = 0
                if (Z_scores_observed[where_is] < 1.96) and (two_in_a_row_Z_lt_196 == 0):
                    two_in_a_row_Z_lt_196 = 1
                    pos_first_ns = where_is
            if (Z_scores_observed[where_is] < 1.96) and (least_dist_past_peak_Z_lt_196 == 0) and (where_is > peak):
                least_dist_past_peak_Z_lt_196 = where_is


###################### new improved case control flow method ###############################

        # case 1 – floor has been found where Z(floor) > 1.96 and Z(floor + 1 km) and Z(floor + 2 km) are both non-significant
        if floor_reached == 1:
            lose_sig_dist_test = floor # value of floor was set earlier when test status loop determined that floor had been reached
            print "maximum range of significance has been found at distance in km = "+str(floor)

        # case 2 – current lose_sig_dist_test value is way below likely final value as all Z-scores are > 10
        elif (consecutive_peak_to_curr_count == peak_to_curr_gt196_count) and (Z_scores_observed[lose_sig_test_dist] > 10):
            lose_sig_test_dist += 12 # big jump out of 12 km from previous target
        
        # case 3 – current lose_sig_dist_test value is below likely final value
        elif (consecutive_peak_to_curr_count == peak_to_curr_gt196_count):
            lose_sig_test_dist += 5

        # case 4 – contiguous dataset from peak to current lose_sig_test_dist value already exists
        elif (contiguity_peak_to_curr == 1):
            lose_sig_test_dist += 5

        # case 5 – two values in range peak to curr position have Z-scores less than 1.96
        elif (two_in_a_row_Z_lt_196 == 2):
            lose_sig_test_dist -= 5 # drop back down to fill in non-contiguous values

        # case 6 – should never hit this condition if coding is correct
        elif (pos_two_in_a_row > lose_sig_test_dist):
            lose_sig_test_dist += 5

        # case 7 – aware of some distance past current test position where Z is non-significant
        elif (least_dist_past_peak_Z_lt_196 > lose_sig_test_dist):
            lose_sig_test_dist += 5

        # case 8 – aware of some distance short of current test position where Z is non-significant
        elif (least_dist_past_peak_Z_lt_196 < lose_sig_test_dist):
            lose_sig_test_dist -= 5

        # all other possibilities seem remote!!!
        else:
            print "how did we ever reach this spot? "

    print "done now with weed "+weed_name
    if escape_count > 200:
        print "trouble at line 196"
    print escape_count
    print "done fini kaput"
    print "maximum range of MI sigificance found at "+str(floor)+" with Z-value = "+str(Z_scores_observed[floor])
    summary_file.write(weed_name+" maximum range of MI significance = "+str(floor)+" kilometers with Z-score = "+str(Z_scores_observed[floor])+"\n")
#    summary_file.write(weed_name+" shortest non-signif. MI distance = "+str(shortest_ns)+" kilometers with Z-score = "+str(Z_scores_observed[shortest_ns])+"\n")
    summary_file.write(str(gp_count)+" gp calls this run. "+str(escape_count)+" = escape_count value"+"\n")
    summary_file.flush()
    MI_results_file.close() 

############# here we try Getis-Ord General G instead of Morans I
    # algorithm to find distance of maximum GG Z-score

    first_try = 0 # do we ever want to use wider test range of first try = 1?
    test_distance = peak # use MI peak distance as initial GG target
    lose_sig_test_dist = floor # use final max sig distance from MI as GG target
    GG_results_name = "C:\\DataFiles\\OSUSeedCert\\SpatialStats\\GG_"+weed_name+"_variable_distance_results_for_max_Z_score.txt"
    GG_results_file = open(GG_results_name,"a+")
    print GG_results_name+" "+weed_name
    print "GG target lose sig test dist from MI floor = "+str(lose_sig_test_dist)
    peak_reached = 0
    floor_reached = 0
    Z_scores_observed = {-1:0.0,0:0.0} # dummy values of Z = 0,0 for distance of influence = -1,0 ## -1 needed to give full range around possible solution of 1000 m
    # now load any existing data from previously written file
    already_made = GG_results_file.readline()
    max_Z = 0
    temp_peak = 0
    while (len(already_made) > 0) and (already_made != "\n"): # skip this if nothing is in file
        #convert line into dictionary entry
        space1 = already_made.find(" ")
        space2 = already_made.find(" ",space1 + 1)
        space3 = already_made.find(" ",space2 + 1)
        space4 = already_made.find(" ",space3 + 1)
        final_separator = already_made.find("; ",space4 + 1)
        existing_distance = int(already_made[space1 + 1:space2])
        existing_zscore = float(already_made[final_separator + 2:])
        if existing_zscore < 0:
            existing_zscore = 0
        Z_scores_observed[existing_distance] = existing_zscore
        # print str(existing_distance)+" "+str(existing_zscore)
        already_made = GG_results_file.readline() # keep reading in more lines until while loop stops
        first_try = 0 # have some results in stored text file
        if existing_zscore > max_Z:
            max_Z = existing_zscore
            temp_peak = existing_distance
    print Z_scores_observed
    print str(max_Z)+" = max Z-score in text file "+str(temp_peak)+" = distance in km"
    GG_results_file.close()
    GG_results_file = open(GG_results_name,"a")
#    if temp_peak > 0:
#        if test_distance == 10: # change to better value if we've already run program for this weed and written data to text file
#            test_distance = temp_peak
#        if test_distance != 10:
#            print "not changing test distance from "+str(test_distance)+" km in local data to "+str(temp_peak)+" in results stored in text file"
#            # leave test_distance alone unless we add more code 
    while peak_reached == 0:
        for this_shift in range(-3,4): # plus and minus 1, 2, and 3 km from test_distance except during first_try on new weed
            if first_try == 0:
                shifting = this_shift
            else:
                shifting = (2 + this_shift) * 10 # generate wide range 10,20,30,40,50 km if no preliminary search for peak
            local_distance = test_distance + shifting
            print "local distance value is "+str(local_distance)
            if local_distance <= 0:
                local_distance = 1 # do not run at less than 1000 m range
            if not Z_scores_observed.has_key(local_distance): # only run if this distance hasn't already been done
                # Local variables...        
                Observed_General_G = "0"
                ZScore = "0"
                GG_results_file.write(weed_name+" "+str(local_distance)+" in kilometers ")  
                param_distance = 3280.839895 * local_distance  # conversion factor is feet in a kilometer    
                # Process: High/Low Clustering (General G)...
                curr_results = gp.HighLowClustering_stats(LinnBoundaries, weed_name, "false", "Fixed Distance Band", "Euclidean Distance", "None", param_distance, "")
                gp_count +=1
                GG_results_file.writelines(curr_results)
                GG_results_file.write("\n")
                GG_results_file.flush()
                myZ = float(curr_results[curr_results.find(";")+1:])
                if myZ < 0:
                    myZ = 0 # do need keep negative Z-scores as they make no sense in our context and might break the search procedure for maximum Z-score
                Z_scores_observed[local_distance] = myZ
        values_so_far = Z_scores_observed.values()
        max_value = max(values_so_far)
        distances_so_far = Z_scores_observed.keys()
        peak = 0
        for where_is_peak in distances_so_far:
            if Z_scores_observed[where_is_peak] == max_value:
                peak = where_is_peak # this logic will get last value of several duplicate maximum values, but duplicates are very unlikely!!!  
        contiguity = 1
        for testing_contiguity in range(-2,3): # testing whether data exists in a contiguous range of -2,-1,+1,+2 around observed peak
            this_distance = peak + testing_contiguity
            if not Z_scores_observed.has_key(this_distance):
                contiguity = 0       
        if contiguity == 1:
            peak_reached = 1
        else: # still haven't found the peak, need to keep moving up or down in distance
            test_distance = peak
            first_try = 0 # had been set to 0 before start of Gen G peak test anyway
            print "new test distance in kilometers = "+str(test_distance)
    print "peak in GG Z-scores found at "+str(peak)+" with Z-value = "+str(Z_scores_observed[peak])
    GGsummary_file.write(weed_name+"                 peak GG distance = "+str(peak)+" kilometers with Z-score = "+str(Z_scores_observed[peak])+"\n")
    GGsummary_file.write(str(gp_count)+" gp calls this run."+"\n")
    GGsummary_file.flush()

    # now trying to find distance for minimum Z-score > 1.96 with adjacent larger distances non-significant and shorter one more significant
    print "now running the find distance at which significance is first lost algorithm"
    new_min = Z_scores_observed[peak] # start new_min as high as any Z-score calculated so far
    if new_min < 1.96:
        print "warning, will robinson, danger!!!! - maximum observed Z-score was less than that required for 5% significance level"
        
    floor_reached = 0 # forgot to give this a default value  
    shortest_found = 0  
    shortest_ns = max(distances_so_far) # a safe default value!
    if shortest_ns < lose_sig_test_dist:
        shortest_ns = lose_sig_test_dist # fixes case of peak search not loading dictionary with distance as large as initial estimate of lose_sig_test_dist

    old_floor = shortest_ns - 1 # new control variable
    stuck = 0

    escape_count = 0
    while floor_reached == 0:
        escape_count += 1
        if escape_count > 200:
            print "trouble at line 141"
            break
        for this_shift in range(-3,4): # plus and minus 1, 2, and 3 km from lose_sig_test_dist
            escape_count += 1
            if escape_count > 200:
                print "trouble at line 143"
                break
            local_distance = lose_sig_test_dist + this_shift # starts with lose_sig_test_dist = 53 for AGRRE
            print "local distance value is "+str(local_distance)
            if local_distance <= 0:
                local_distance = 1 # do not run at less than 1000 m range
            if not Z_scores_observed.has_key(local_distance): # only run if this distance hasn't already been done
                print "in the floor reach -3,-2,-1,0,1,2,3 loop"
                # Local variables...        
                Observed_General_G = "0"
                ZScore = "0"
                GG_results_file.write(weed_name+" "+str(local_distance)+" in kilometers ")  
                param_distance = 3280.839895 * local_distance  # conversion factor is feet in a kilometer    
                # Process: High/Low Clustering (General G)...
                curr_results = gp.HighLowClustering_stats(LinnBoundaries, weed_name, "false", "Fixed Distance Band", "Euclidean Distance", "None", param_distance, "")
                gp_count += 1
                GG_results_file.writelines(curr_results)
                GG_results_file.write("\n")
                GG_results_file.flush()
                myZ = float(curr_results[curr_results.find(";")+1:])
                if myZ < 0:
                    myZ = 0 # do need keep negative Z-scores as they make no sense in our context and might break the search procedure for maximum Z-score
                Z_scores_observed[local_distance] = myZ
        values_so_far = Z_scores_observed.values()
        values_so_far.sort()
        print values_so_far # dubugging printout
        old_min = new_min
        #for temp_Z in values_so_far: # need to track non-zero minimums
            # need to limit next test to just cases where distance > peak
        #    if (temp_Z < new_min) and (temp_Z >= 1.96): # using loss of 5% significance level for floor in finding maximum range of spatial autocorrelation
        #        new_min = temp_Z

     

############################# status condition ########################

        distances_so_far = Z_scores_observed.keys()
        testing_ns = max(distances_so_far) 
        consecutive_peak_to_curr_count = 0    
        peak_to_curr_gt196_count = 0   
        contiguity_peak_to_curr = 1 
        two_in_a_row_Z_lt_196 = 0
        pos_first_ns = testing_ns + 1 # safe default value
        pos_two_in_a_row = 0 # safe default
        least_dist_past_peak_Z_lt_196 = 0 # safe default
        distances_so_far.sort() 

        for test_contiguity in range(peak,lose_sig_test_dist+1):
            if not(Z_scores_observed.has_key(test_contiguity)):
                contiguity_peak_to_curr = 0 # contiguity does not yet exist in range peak to lose_sig_test_dist

        for where_is in distances_so_far: # loop thru distances in order, focusing on conditions past the peak
            if Z_scores_observed.has_key(where_is + 1) and Z_scores_observed.has_key(where_is + 2): # verify data exists at next two step on out
                if (Z_scores_observed[where_is] > 1.96) and (Z_scores_observed[where_is + 1] < 1.96) and (Z_scores_observed[where_is + 2] < 1.96) and (floor_reached == 0) and (where_is >= peak):
                    floor_reached = 1
                    floor = where_is # this completes things that needed to be done once floor location has been found

            if (where_is > peak) and (where_is <= lose_sig_test_dist):
                consecutive_peak_to_curr_count += 1
                if (Z_scores_observed[where_is] > 1.96):
                    peak_to_curr_gt196_count += 1
                if (Z_scores_observed[where_is] < 1.96) and (two_in_a_row_Z_lt_196 == 1) and (where_is == pos_first_ns + 1):
                    two_in_a_row_Z_lt_196 = 2
                    pos_two_in_a_row = where_is
                    # this is trigger for Case 5
                else: # reset count of two in a row NS Z-scores to 0
                    if (two_in_a_row_Z_lt_196 == 1):
                        two_in_a_row_Z_lt_196 = 0
                if (Z_scores_observed[where_is] < 1.96) and (two_in_a_row_Z_lt_196 == 0):
                    two_in_a_row_Z_lt_196 = 1
                    pos_first_ns = where_is
            if (Z_scores_observed[where_is] < 1.96) and (least_dist_past_peak_Z_lt_196 == 0) and (where_is > peak):
                least_dist_past_peak_Z_lt_196 = where_is


###################### new improved case control flow method ###############################

        # case 1 – floor has been found where Z(floor) > 1.96 and Z(floor + 1 km) and Z(floor + 2 km) are both non-significant
        if floor_reached == 1:
            lose_sig_dist_test = floor # value of floor was set earlier when test status loop determined that floor had been reached
            print "maximum range of significance has been found at distance in km = "+str(floor)

        # case 2 – current lose_sig_dist_test value is way below likely final value as all Z-scores are > 10
        elif (consecutive_peak_to_curr_count == peak_to_curr_gt196_count) and (Z_scores_observed[lose_sig_test_dist] > 10):
            lose_sig_test_dist += 12 # big jump out of 12 km from previous target
        
        # case 3 – current lose_sig_dist_test value is below likely final value
        elif (consecutive_peak_to_curr_count == peak_to_curr_gt196_count):
            lose_sig_test_dist += 5

        # case 4 – contiguous dataset from peak to current lose_sig_test_dist value already exists
        elif (contiguity_peak_to_curr == 1):
            lose_sig_test_dist += 5

        # case 5 – two values in range peak to curr position have Z-scores less than 1.96
        elif (two_in_a_row_Z_lt_196 == 2):
            lose_sig_test_dist -= 5 # drop back down to fill in non-contiguous values

        # case 6 – should never hit this condition if coding is correct
        elif (pos_two_in_a_row > lose_sig_test_dist):
            lose_sig_test_dist += 5

        # case 7 – aware of some distance past current test position where Z is non-significant
        elif (least_dist_past_peak_Z_lt_196 > lose_sig_test_dist):
            lose_sig_test_dist += 5

        # case 8 – aware of some distance short of current test position where Z is non-significant
        elif (least_dist_past_peak_Z_lt_196 < lose_sig_test_dist):
            lose_sig_test_dist -= 5

        # all other possibilities seem remote!!!
        else:
            print "how did we ever reach this spot? "
       
    print "done now with weed "+weed_name
    if escape_count > 200:
        print "trouble at line 196"
    print escape_count
    print "done fini kaput"
    print "maximum range of GG sigificance found at "+str(floor)+" with Z-value = "+str(Z_scores_observed[floor])
    GGsummary_file.write(weed_name+" maximum range of GG significance = "+str(floor)+" kilometers with Z-score = "+str(Z_scores_observed[floor])+"\n")
    #GGsummary_file.write(weed_name+" shortest non-signif. GG distance = "+str(shortest_ns)+" kilometers with Z-score = "+str(Z_scores_observed[shortest_ns])+"\n")
    #GGsummary_file.write(weed_name+" true_max non-signif. GG distance = "+str(true_max)+" kilometers with Z-score = "+str(Z_scores_observed[true_max])+"\n")
    GGsummary_file.write(str(gp_count)+" gp calls this run. "+str(escape_count)+" = escape_count value"+"\n")
    GGsummary_file.flush()
    GG_results_file.close() 
    
summary_file.close()
GGsummary_file.close()