Dateien nach "ILP/Vanilla" hochladen

ILP/Vanilla/nmrSimilarityVanillien.py

@@ -49,7 +49,7 @@ CMCOUMARICACID = {
     4: ([141.29], [1]),
     5: ([121.09], [1]),
     6: ([134.46], [1]),
-    7: ([121.09], [1]),
+    7: ([121.69], [1]),
     8: ([164.15], [1]),
     9: ([122.36], [1]),
 }
@@ -255,26 +255,42 @@ def correction(spectra, corretionppm):
         newspectra[id] = ([adjustedshift], height)
     return newspectra
 
+def addspectra(spectrum1, spectrum2):
+    spectrum = spectrum1.copy()
+    for _, ([ppm2], [height2]) in spectrum2.items():
+        for peak, ([ppm1], [height1]) in spectrum1.items():
+            if ppm1 == ppm2:
+                spectrum[peak] == ([ppm1], [height1 + height2])
+                continue
+        spectrum[len(spectrum) + 1] = ([ppm2], [height2])
+    return(spectrum)
+
 def main():
-    spectrumref = CNMR2
-    #1H-NMR Spectra ignoriert, da meiste H sauer, da an N gebunden
-    #spectra = [HCINNAMICACID, HPCOUMARICACID, HMCOUMARICACID, HBENZALDEHYD, HCAFFEICACID, H3HYDROXYBENZALDEHYD, H4HYDROXYBENZALDEHYD, H34DIHYDROXYBENZALDEHYD]
-    spectra = [CCINNAMICACID, CPCOUMARICACID, CMCOUMARICACID, CBENZALDEHYD, CCAFFEICACID, C3HYDROXYBENZALDEHYD, C4HYDROXYBENZALDEHYD, C34DIHYDROXYBENZALDEHYD]
-    spectranames = ["CINNAMICACID", "PCOUMARICACID", "MCOUMARICACID", "BENZALDEHYD", "CAFFEICACID", "3HYDROXYBENZALDEHYD", "4HYDROXYBENZALDEHYD", "34DIHYDROXYBENZALDEHYD"]
-    likelihood = []
-    for spectrumtrue in spectra:
-        similaritybycorrection = []
-        #Paper Chemical reviews Carbons bound to Heavy atoms (TMS) to high -> this could be reason for too high values.
-        correctionvalues = [2.5] #np.arange(0, 1.51, 0.01) #for C tested np.arange(-0.37, 7.64, 0.1) 0, 2.63, 1 (for CNMR3), 1.5 (true for all ref, 8.37 + 1.5 for the precision), 1 (good for first, ok for second, third because only 7 better/equal but for first much higher) np.arange(-1.5, 1.49, 0.01) (only for first), for H 0 (not good), np.arange(-0.26, 0.34, 0.01), -0.06 for first, second never first either 17 or caf higher np.arange(-0.15, 0.16, 0.01) good measure
-        for correctionvalue in correctionvalues: 
-            spectrumrefcorrected = correction(spectrumref, correctionvalue) #CCAFFEINE 11 (klappt hier sehr gut) CCAFFEINE2 12 CPARAXANTHINE 10 CNMR1 9, 10 o 11 (sehr gut) CNMR2 10 o 11
-            similaritylist = []
-            binwidthlist = np.arange(0.1, 3.9, 0.1)
-            for i in binwidthlist:
-                similaritylist.append(similarity_nmr(spectrumtrue, spectrumrefcorrected, i))
-            similaritymean = sum(similaritylist) / len(similaritylist)
-            similaritybycorrection.append(similaritymean)
-        likelihood.append(round(sum(similaritybycorrection)/len(similaritybycorrection), 2))
-    print(likelihood)
+    normalize = False
+    spectrumrefs = [CNMR1, CNMR2]
+    for spectrumref in spectrumrefs:
+        #spectra = [HCINNAMICACID, HPCOUMARICACID, HMCOUMARICACID, HBENZALDEHYD, HCAFFEICACID, H3HYDROXYBENZALDEHYD, H4HYDROXYBENZALDEHYD, H34DIHYDROXYBENZALDEHYD]
+        spectra = [CCINNAMICACID, CPCOUMARICACID, CMCOUMARICACID, CBENZALDEHYD, CCAFFEICACID, C3HYDROXYBENZALDEHYD, C4HYDROXYBENZALDEHYD, C34DIHYDROXYBENZALDEHYD]
+        spectranames = ["CINNAMICACID", "PCOUMARICACID", "MCOUMARICACID", "BENZALDEHYD", "CAFFEICACID", "3HYDROXYBENZALDEHYD", "4HYDROXYBENZALDEHYD", "34DIHYDROXYBENZALDEHYD"]
+        likelihoods = []
+        for spectrumtrue in spectra:
+            similaritybycorrection = []
+            #Paper Chemical reviews Carbons bound to Heavy atoms (TMS) to high -> this could be reason for too high values.
+            correctionvalues = [2.5] #np.arange(0, 1.51, 0.01) #for C tested np.arange(-0.37, 7.64, 0.1) 0, 2.63, 1 (for CNMR3), 1.5 (true for all ref, 8.37 + 1.5 for the precision), 1 (good for first, ok for second, third because only 7 better/equal but for first much higher) np.arange(-1.5, 1.49, 0.01) (only for first), for H 0 (not good), np.arange(-0.26, 0.34, 0.01), -0.06 for first, second never first either 17 or caf higher np.arange(-0.15, 0.16, 0.01) good measure
+            for correctionvalue in correctionvalues: 
+                spectrumrefcorrected = correction(spectrumref, correctionvalue) #CCAFFEINE 11 (klappt hier sehr gut) CCAFFEINE2 12 CPARAXANTHINE 10 CNMR1 9, 10 o 11 (sehr gut) CNMR2 10 o 11
+                similaritylist = []
+                binwidthlist = np.arange(0.1, 3.9, 0.1)
+                for i in binwidthlist:
+                    similaritylist.append(similarity_nmr(spectrumtrue, spectrumrefcorrected, i))
+                similaritymean = sum(similaritylist) / len(similaritylist)
+                similaritybycorrection.append(similaritymean)
+            likelihoods.append(sum(similaritybycorrection)/len(similaritybycorrection))
+    if normalize:
+        normalizedlikelihood = [round(likelihood/np.sum(likelihoods), 2) for likelihood in likelihoods]
+        print(normalizedlikelihood)
+    if not normalize:
+        notnormalizedlikelihood = [round(likelihood, 2) for likelihood in likelihoods]
+        print(notnormalizedlikelihood)
 if __name__ == "__main__":
     main()